Novel Method to Predict Non-Measured Parameters of Industrial Gas Turbines Using Kalman Filter Technique

Gas turbines are most widely used for power generation and operate under various conditions and loads. Gas turbine control is important to cope with various situations, and the turbine inlet temperature (TIT) is the most important parameter because it is directly related to the power output and life cycle of the turbine. Thus, precise prediction and control of the TIT are important in terms of the stable operation and life cycle management of gas turbines. This paper proposes a new method to predict non-measured parameters such as the air flow and TIT using Kalman filter techniques. The Kalman filter is widely used for estimating the instantaneous state of a system and can estimate non-measured parameters. The Kalman filter algorithm was implemented in a gas turbine analysis program using MATLAB. The reliability of the new method was verified through various case studies using virtual data and real operating data. The results were compared with those of a model-based gas turbine diagnostics program. The computing time of the Kalman filter and model-based diagnostics program were also compared to confirm the capability of the new method. The results indicate that the new method is more suitable for diagnostics and monitoring applications than the model-based analysis program. Finally, two case studies were performed to confirm the feasibility of the new method using two virtual datasets. The results confirm that the Kalman filter can predict the non-measured parameters precisely.