Turbine system identification: experimental results

A series of tests were conducted on a land based GE gas turbine to determine its dynamic characteristics. At each operating point, models were evaluated on the basis of how well they fit two different sets of data. The data differed in the way the excitation signal entered the system. The objective was to obtain an accurate transient model of the turbine that could be used for controller design. The results show that least squares identification is an effective method for accurately modeling gas turbine systems. This paper gives an brief overview of the results