Data reconciliation-based simulation of thermal power plants for performance estimation and digital twin development

Abstract

High-fidelity simulation of thermal systems remains challenging due to online measurement errors and inherent mismatches of thermal system models. Previous model-based simulation methods always cause certain deviations between simulated data and online measured data. This paper proposes a simulation framework to solve the problem by making full use of thermal system model and online measurements. First, a grey-box thermal system model is established. Second, key performance parameters obtained from the model are regarded as virtual measurements to be reconciled with online measurements. An in-service double reheat power plant is selected as a case study. Results show that the proposed simulation framework alleviates the error accumulation and decreases the simulation error from 0.78 percent to 0.13 percent. Moreover, performance estimation uncertainties of the thermal system and key components are reduced by 0.51–86.68 percent. This provides a new concept and potential high-fidelity simulation framework to develop digital twins for thermal power plants.