This page outlines the PV modelling methodology and specifications of Solcast's Advanced PV Model, through which users can access forecasts and modelled actuals with global-coverage across Live (-7 days to present moment), and Forecast (present moment to +14 days) time periods.
Model overview
The Solcast Advanced PV Model, which produces the data parameter Advanced PV Power Output (MW) is designed for PV plants where system specifications are known, and where accuracy of production estimates and forecasts are important. Solcast's Advanced PV model is based on the open source pvlib-python library, with several proprietary extensions added by the Solcast team.
To use the Advanced PV model, a user must first create their site(s) in the Solcast system, either manually via the Solcast API Toolkit web portal or programmatically via the Solcast API. These sites exist to store, list, and edit the PV plant specifications used by the model, and site location information. There is no delay between site creation and availability of data, due to Solcast's real-time global 1-2km resolution data capability. Each site is identified with a unique site resource ID.
The model's output data parameter Advanced PV Power Output (MW) (Solcast API parameter pv_power_advanced) is returned by the Solcast API when a user request to the API is made with the site's resource ID. These PV power values represent Solcast's irradiance and weather inputs (including satellite cloud tracking and forecasting), and the user-specified PV plant specifications. In the case of forecasts, which span from the present time up to 14 days ahead, probabilistic 10th and 90th percentile values are also returned by the API as pv_power_advanced10 and pv_power_advanced90.
Input data and model chain
The following flowchart diagram summarizes the input data and model chain used in Solcast's Advanced PV model. Orange boxes represent inputs (plant specifications metadata; and time-dynamic weather data inputs), blue boxes represent key steps in the model chain, and the final box represents the model output (i.e. the Solcast data parameter Advanced PV Power Output (MW).
The Advanced PV model uses Solcast's irradiance and weather data as its time-dynamic input. The Solcast data parameters used by the model include irradiance parameters Direct Normal Irradiance (DNI) and Diffuse Horizontal Irradiance (DHI), ambient parameters Air Temperature and Wind Speed (10m), satellite-derived surface reflectivity parameter Albedo Daily, and a dynamic site-specific Snow Soiling Loss – Ground Mounted based on irradiance, temperature and precipitation. For more information about this input data, please refer to Solcast Irradiance and Weather Data Inputs and Algorithms documentation.
The model uses user-configurable plant specifications to generate a site-specific physical power simulation. Plant specifications utilised by the model include array geometry and tracking type, module and inverter information, horizon shading, dust soiling losses, other losses including degradation, presence of tracking algorithms, and bi-facial module support. A full list of the plant specifications used by the model is provided in Section 4 of this document. The model can run on a simplified set of basic plant specifications in cases where full specifications are not known. The minimum set of specifications required include location, AC capacity, and tracking type. In these circumstances, Solcast makes reasonable estimates of the remaining parameters based on the entered values. However, in general users should be aware that the input of more metadata parameters will increase both the accuracy and precision of the physical power simulation. If detailed specifications cannot be found, Solcast recommends tuning with production measurement data, which we offer as a service, described in Section 3 of this document.
The model chain is based on the open source software package pvlib-python, with a number of proprietary extensions built and maintained by the Solcast team. For more information on pvlib-python, see the pvlib python documentation.
Improving forecast accuracy using historical production data
The Solcast Advanced PV Model is primarily designed for accurate forecasting of power output and energy production for minutes to weeks ahead. The 20+ physical PV plant specifications used by the model are primarily a means of achieving this forecast accuracy. Where a long history (at least three weeks) of production measurement data of sufficient quality is available, the initially-entered and estimated/default values of these PV plant specification parameters can be reviewed and refined empirically with the objective of minimising forecast error.
Solcast offers such a service to its Enterprise customers. To make a request, you will need to provide all known plant specifications, PVsyst reports where available, and a time series of power output or energy production at hourly granularity or finer, of at least three weeks duration (ideally 12 months, and ideally ending close to the present date). The ideal time granularity is 5 or 15 minutes. If available, please also provide measured values of GHI and GTI, measured values of total plant availability (%), and flags for periods where the plant was constrained or curtailed.
Appendix of PV plant specification paramaters
The PV plant specification parameters of the Solcast Advanced PV Model are listed in the following set of tables. For most parameters (all except those denoted with “Required field”), Solcast will pre-fill these plant specification parameters with default values based on the above basic specification parameters. However, accepting the defaults could still lead to large errors, so you are encouraged to review these assumptions and replace them with actual plant specifications. You will be able to edit these settings later.
| Name | Definition | How we use this |
|---|---|---|
| Name | The user-defined name of the site | Used to display a name for your site. Can be changed |
| Latitude (site centre) | Site latitude (centre of site) in decimal degrees (EPSG:4326). | Used to locate your site. Cannot be changed |
| Longitude (site centre) | Site longitude (centre of site) in decimal degrees (EPSG:4326). | Used to locate your site. Cannot be changed |
| Tracking type | The type of sun-tracking or geometrical configuration of your site's modules | Used to calculate the incident irradiance for your modules |
| Install date | The date when your site was installed | Used to derate your module (DC) production gradually with age, at a rate dependent on your Module Type |
| AC capacity (inverters) in MW | Total inverter (nameplate) capacity in MW. This is the highest potential output of the system before any Site Export Limit is applied | Used to model the conversion of DC power to AC by your inverters |
| DC capacity (modules) in MW | Total module capacity in MW. Usually slightly higher than the AC capacity. | Used to model the generation of DC power by your modules |
| Grid export limit in MW | The maximum power export limit in MW that is allowed by the site's connection with the network operator | Used to place a final cap on your AC power output. Only impacts your AC power if the grid export limit is set lower than the AC capacity |
| Name | Definition | How we use this |
|---|---|---|
| Module type | The type of material or technology used in your site's PV modules. | Used to estimate your module temperature derating coefficient (unless you specify your own coefficient) and used to estimate your module age derating |
| Ground coverage ratio | The proportion of the site's ground area covered by modules. | Used to calculate the incident irradiance for your modules |
| Module temperature derating coefficient | The factor by which your site's module (DC) production will be derated with increasing temperature | Used to derate your module (DC) production gradually with increasing temperature |
| Peak inverter efficiency | The peak efficiency value in your inverter efficiency curve | Used to scale the conversion efficiency of DC to AC, as a function of the inverter load |
| Age degradation derating coefficient | The factor by which the whole system will be derated per year since the Install Date | Used to calculate time dependent system los |
| Other losses derating coefficient | The factor by which the whole system will be derated due to causes other than temperature and degradation | Used to calculate other system losses including mismatch, wiring and connection |
| Name | Definition | How we use this |
|---|---|---|
| Module tilt angle | The off-horizontal tilt angle of modules for a fixed-tilt site. | Used to calculate the incident irradiance for your modules |
| Module azimuth angle | The off-north-facing direction on the horizon in which the modules are facing for a fixed-tilt site. | Used to calculate the incident irradiance for your modules |
| Name | Definition | How we use this |
|---|---|---|
| Tracker axis azimuth angle | The off north-south azimuth angle for a horizontal single axis tracking site. Most commonly this will be close to zero | Used to calculate the incident irradiance for your modules |
| Tracker maximum rotation angle | The maximum off-horizontal angle for a horizontal single axis tracking site. | Used to calculate the incident irradiance for your modules |
| Tracker back-tracking | Whether the trackers backtrack at low solar elevation angles, for a horizontal single axis tracking site. | Used to calculate the incident irradiance for your modules |
| Tracker smart tracking | Whether the trackers move to horizontal during cloudy periods with zero DNI, for a horizontal single axis tracking site. | Used to calculate the incident irradiance for your modules |
| Name | Definition | How we use this |
|---|---|---|
| Average terrain slope | The average terrain slope in degrees of your site. A site with no terrain slope has a value of zero. | Used to calculate the incident irradiance for your modules |
| Average terrain azimuth | The average terrain slope downhill direction. | Used to calculate the incident irradiance for your modules. Terrain Azimuth is meaningful only when terrain_slope>0. |
| Average dust soiling losses | The average proportion of module production lost due to dust soiling, from 0 to 1. E.g. 0.02 for April means a 2% soiling loss will be applied throughout April. The value entered should reflect the impact of cleaning activity at your site. | Used to calculate the module (DC) production We accept either a single yearly average value, or a set of 12 monthly average values. |
| Topographic horizon elevation | The elevation angle of topography (e.g. hills and mountains) surrounding the general area of your site, at a set of specified different azimuth (direction) angles. | Used to calculate the incident irradiance for your modules |
Appendix of "apply" plant condition parameters
These optional parameters exist to accommodate deviations in the actual operating condition of the plant, which can vary from minute-to-minute or day-to-day and therefore are not suitable to be semi-permanent plant specification metadata like the parameters above. When these parameters are not used, the above plant specification metadata is used. The "apply" parameters can be optionally specified in the API request, with no delay in receiving forecasts back from the Solcast API. Not all "apply" parameters need to be provided in the request, only the parameters relating to the actual changes in plant operating conditions.
| Name | Definition | How we use this |
|---|---|---|
| apply_availability | Proportion of the site's total AC (inverter) capacity that is currently generating or expected to be generating during the forecast request period. E.g. if you specify a 30% availability, your returned power will be a maximum of 30% of AC capacity | To adjust the pv power value returned by the Solcast API. |
| apply_constraint | Constraint on site's total AC production, applied as a cap in the same way as the metadata parameter Site Export Limit. This will constrain all Solcast power values to be no higher than the apply_constraint value you specify. If you need an unconstrained forecast, you should not use this parameter. | To adjust the pv power value returned by the Solcast API. |
| apply_dust_soiling | A user-override for dust_soiling_average. If you specify this parameter in your API call, we will replace the annual or monthly average dust soiling values for the site with the value you specify in your API call. E.g. if you specify a 0.7 dust soiling, your returned power will be reduced by 70%. | Used to calculate the module (DC) production |
| apply_snow_soiling | A user-override for Solcast's dynamic snow soiling, which is based on global snow cover and weather forecast data, and changes from hour-to-hour. If you specify this parameter in your API call (e.g. if snow clearing has just been performed), we will replace the Solcast dynamic hour-to-hour value with the single value that you specify. E.g. if you specify a 0.7 snow soiling, your returned power will be reduced by 70%. | Used to calculate the module (DC) production |
| apply_tracker_inactive | Indicating if trackers are inactive. If True, panels are assumed all facing up (i.e. zero rotation). Only has effect if your site has a tracking_type that is not "fixed" | Used to calculate the module (DC) production |