Building physically plausible digital twins is the key to optimize power plants.
Read here how an expert consultancy used Visplore to reduce the time to validate digital twins by 20% – and saved around $10,000 of cost per project for data preparation and model adjustment.
Visplore features pre-configured cockpits for flexible self-service data exploration, data preparation and model validation.
Quickly grasp the error distribution including trends and outliers of forecast models over time.
Click Image to enlarge
Violation of expected value from digital twin.
Click Image to enlarge
An increasing number of power plants are operated with partial load, being tied to consumption or complementing renewables. However, since historical data is often not readily available for partial load, energy generation companies struggle to build reliable, realistic digital twins for optimizing their power plants. Challenges like missing representative data points, or data quality issues make purely data-driven machine learning approaches hard. Therefore, it is necessary that domain experts themselves are able to explore and validate a power plant’s technical data to find those time periods or situations where deviations of the models to the actual physical conditions are significant.
Visplore empowers domain experts to explore and clean hundreds of complex time series and accelerate the digital twins validation by 20%, on average.
