Adaptive predictive control of combustor NO_x emissions

摘要

The design and implementation for the active control of oxides of nitrogen (NO_x) emissions from gas turbine combustors is considered in this paper. Predictive control techniques with both fixed and adaptive parameters are employed. An on-line parameter estimation algorithm provides good modelling of the nonlinear characteristics of the combustor NO_x process. To obtain a compari- son with other control techniques, an optimal PID controller is detailed. The actuator that controls the gas flow to the combustor is an electromagnetic shaker. Gain-scheduling control is applied to provide accurate position tracking of the shaker whose character- istics vary between different operating points. These control strategies are implemented using the SIMULINK/dSPACE controller development environment. Their performance is evaluated on an atmospheric test rig fitted with a commercial combustor.