Reliability Evaluation Through Analytics and Data for Emerging Technology in Photovoltaics

TNO – Digital Twin of PV Devices

Virtual Access

  • TNO builds a digital twin of a photovoltaic device based on collected data
  • The digital twin is used to predict device characteristics of outdoor devices under a variety of circumstances
  • Beyond performance prediction, the tool can also identify degradation effects
Digital Twinning Methodology
  • Involves a set of linked computation sets:
    • Set 1: computes device performance from outdoor data
    • Set 2: computes device characteristics from a one-diode model
  • When only the one-diode model is applied, results are typically less satisfactory
Modeling Enhancement
  • A multi-linear regression model is used to improve results:
    • Applicable to single-junction devices
    • Further expanded to support multi-junction devices
  • For multi-junction devices, outdoor data must include spectral information

Users can request access to TNO’s AI-based digital twin tool; upon approval, TNO explains the model and required input data (type, volume, quality, structure). Data management may use NOMAD to simplify exchange with other users. The digital twin is built for a single device structure, based on data from up to 4 similar PV devices, with input data structured for model training.

Provided on a best-effort basis — satisfactory results aren’t guaranteed, so users are encouraged to contact TNO before applying. Access to TNO’s outdoor setup is also available, leveraging extensive experience combining outdoor data with digital-twin modelling.