Case Studies /
Analytics & Financing Utility-scale Solar Park
A globally leading PV module manufacturer and plant developer ran into difficulties when building a utility scale PV park with their newest modules. Because the performance and cash flows of the newly developed modules were not known for a timeframe relevant to financing, no lender would provide financing to the project.
Parameters that could not be determined using conventional techniques were forecasted using state-of-the-art artificial intelligence. Among others, life expectancy of the modules, shadowing effects and cell degradation were predicted for the next 20 years.
Based on the determined modules and plants characteristics, a financial model was built to make the whole project “bankable”. By increasing the debt/equity-ratio of this CAPEX-dominated PV-project, a substantial increase in profitability was achieved.
Forecast Grid Imbalance
A utilities dispatch desk is using their conventional generation fleet to balance their renewable and consumption portfolio. They increase conventional output if their portfolio and the grid is undersupplied and decrease if oversupplied. However, taking into account relatively high startup costs of thermal power plants, this decision requires knowledge of the duration of the grid situation.
Using autoregressive and fundamental modelling techniques, the area control error was forecasted for a horizon of six hours in quarter-hourly resolution. Based on this, imbalance settlement prices can be foreseen.
We were able to reach very high precision on the direction, size and settlement price of grid imbalances. All this was achieved using mostly publicly available data. Using the forecast operatively, decisions to start or stop a power plant can be taken with great commercial confidence.
An automotive supplier has 220 power consumers (plant components/machines) in his operations. A project is initiated to cut his electricity spending of currently €2.5m per year.
A traditional measurement approach is discarded due to its high costs. Instead, existing process data is leveraged with our AI to gain detailed knowledge about electricity consumption of every power consumer in every operating mode, thus enabling massive optimization in short time.
Based on the findings, 8 consumers were retrofitted or replaced, 19 consumers were equipped with automatic switches to eliminate standby losses and 2 consumers that were mostly unused were manually disconnected. The resulting savings in electricity spending are 130.000€ per year.