Utilising a digital twin for the troubleshooting of a geothermal power station

Abstract

A digital twin of the Rotokawa power station is under development. The quasi-empirical model is intended to replicate the behaviour of the real plant, based on process engineering first principles. This paper highlights how an initial version of the model was successfully used to troubleshoot a performance decline in the plant’s Organic Rankine Cycle (ORC) bottoming units. Calculation of key plant parameters highlighted potential causes of this decline, and the initial model could then be used to quantify the impacts of different changes. The study results have been used to inform and direct a campaign to restore plant performance.