PlantPredict is a cloud-based web application that allows users to develop solar energy estimates for utility scale PV applications. PlantPredict is specifically tailored to provide monitoring points that align with utility scale power plant standards. A distinguishing feature of PlantPredict is that the weather input is not limited to hourly data; sub hourly weather measurements from on-site meteorological stations can also be used for energy predictions. The power plant builder is a nested hierarchy of subassemblies allowing for nearly identical replication of the as-built site.
PlantPredict is superior to conventional modeling tools in large part because of its cloud based architecture, giving it more flexibility and security than typical desktop software applications. Users can share their data securely within their company’s PlantPredict community while maintaining peace of mind that their data is not spread across several desktop computers. Additionally the cloud environment makes PlantPredict’s API- based architecture possible. To this end, PlantPredict services can be called via a collection of web APIs. This structure allows third party applications to integrate with the API engine and run it remotely. From an Operation and Maintenance perspective, this offers a unique opportunity to automate reporting of actual daily energy generation compared with contractual models.
One key differentiator of PlantPredict is the comprehensive validation testing against both actual plant data and comparisons against other industry applications. Primarily, the benchmarking of 1 GWdc spread across 1GW DC capacity over 20 sites PV systems confirmed PlantPredict under-predicted energy by 0.41% +/- 2.01% (1-sigma). While a validation exercise against PVsyst using 51 equivalent simulations showed a mean energy yield difference of 0.13% +/- 0.52% (1-sigma). These validation exercises demonstrate that PlantPredict is not only highly accurate but able to reproduce results from predictions previously run using other commercially available software. The full algorithm suite is available in the PlantPredict Resource Center to eliminate any “black-box” concerns around how the algorithms are implemented.