An adaptive optimal combustion control strategy

Abstract A strategy for optimizing the performance of a laboratory-scale combustor with respect to volumetric heat release (maximize) and pressure fluctuations (minimize) has been developed. This strategy utilizes (1) actuation techniques that simultaneously control volumetric heat release and pressure fluctuations; (2) sensing techniques that measure combustor performance; and (3) an adaptive optimal control strategy. Actuation techniques were chosen for their ability to alter the unsteady flow associated with the shear layer and recirculating flow regions of the combustor. Periodic spanwise forcing of the inlet boundary layer is used to reduce combustion-induced pressure fluctuations. Crossflow jets upstream of the inlet are used to control volumetric heat release. Sensing techniques were selected to measure or estimate the two performance parameters to be controlled. A fast-response piezoelectric pressure transducer measures the magnitude of the pressure fluctuations. A streamwise array of photodiodes measures light emission from the flame and enables estimation of volumetric heat release. Combustor performance is explicitly defined in terms of a cost function that is a weighted combination of rms pressure fluctuations and mean volumetric heat release. The control strategy performs an on-line minimization of the cost function by continuously seeking the optimal combination of actuator settings and subsequently maintaining the cost at a minimum when the combustor is subject to unknown inlet condition changes, such as flow disturbances. The strategy has been tested with large flow disturbances and found capable of indirectly sensing a change in combustor inlet conditions and finding the new optimal actuator settings.