A data noise reduction model for a direct-fired fuel cell turbine hybrid power system was evaluated using a hardware-based simulation of an integrated gasifier/fuel cell/turbine hybrid cycle (IGFC), implemented through the Hybrid Performance (HyPer) project at the National Energy Technology Laboratory, U.S. Department of Energy (NETL). The Hyper facility is designed to explore dynamic operation of hybrid systems and quantitatively characterize such transient behavior.
The system is controlled by an embedded real-time control platform provided by Woodward Industrial Control. Every sensor is monitored by the platform, and an overall strategy drives the system from start-up to shut-down. Fuel is regulated by a valve which reacts based on the speed of the turbine. There are three optical encoder sensors which are used to monitor turbine speed, and the average of these three sensors is used as feedback for a PID controller, which works to regulate fuel consumption to the combustor. The turbine speed has demonstrated fluctuation in certain conditions, which may be a result of data noise combined with a systemic instability in the flow to the turbine.
This research introduces the method of Double Exponential Smoothing as it is applied to data noise reduction in an embedded control platform. An experimental test was conducted to evaluate the performance of the fuel valve speed control when filtered by a real-time Double Exponential Smoothing Algorithm. The results demonstrate, that when compared with traditional filtering techniques, Double Exponential Smoothing offers a significant improvement in both signal volatility and data latency.
