Application of unscented Kalman filter for condition monitoring of an organic Rankine cycle turbogenerator

This work relates to a project focusing on energy optimization on offshore facilities. On oil and gas platforms it is common practice to employ gas turbines for power production. So as to increase the system performance and reduce emissions, a bottoming cycle unit can be designed with particular emphasis on compactness and reliability.

In such context, organic Rankine cycle turbogenerators are a promising technology. The implementation of an organic Rankine cycle unit is thus considered for the power system of the Draugen offshore platform in the northern sea, which is the case study for this project. Considering the plant dynamics, it is of paramount importance to monitor the peak temperatures within the once-through boiler serving the bottoming unit to prevent the decomposition of the working fluid.

This paper accordingly aims at applying the unscented Kalman filter to estimate the temperature distribution inside the primary heat exchanger by engaging a detailed and distributed model of the system and available measurements. Simulation results prove the robustness of the unscented Kalman filter with respect to process noise, measurement disturbances and initial conditions.