The market requirements with regard to transient operation capabilities of gas turbines (GT) in utility use are becoming stringent. Besides normal frequency support features, gas turbines in local electrical grids are often required to maintain the grid frequency under various situations, including emergencies, such as, loss of national grid connection or trip of a large consumer, etc. These requirements demand high performance and stability of GT control. On the other hand, the environmental aspects are becoming increasingly a public concern. In the past decades, remarkable progress has been made in combustion technologies of heavy-duty gas turbines. Lean premixing is a preferred technology for NOx emission reduction. Because of its flashback and extinction limits, a premix flame has usually a much narrower operation range compared to a diffusional one, adding tight constraint on GT control. This paper demonstrates a successful implementation of a model-based predictor, a proven control technique, in the closed loop control of the ALSTOM GT11N2 gas turbine. First, an online GT model, which is integrated into the GT control algorithm, was developed. By applying appropriate assumption and simplification, this model is capable of simulating the GT process over the whole load operation range with high dynamic accuracy. Secondly, a model-based predictor for accelerating slow measured signals was implemented. It dynamically compensates the system delays in the GT process and in the measuring instruments. Thirdly, the predictor was applied to the GT core control by replacing the measured signals with the accelerated signals. The original control structure was kept unchanged. In order to verify its performance and stability, the new control technique was validated on a real engine. Successful engine tests proved that the model-based predictor improves GT transient operation capabilities.
Rotordynamic Modeling of an Actively Controlled Magnetic Bearing Gas Turbine Engine