Real-Time-Optimization of Jet Mixing via Evolution Strategies

摘要

This paper presents initial findings in the application of evolution strategies lor real-time optimization of jet mixing enhancement. The experimental apparatus consists of a small-scale jet turbine and pulsed fluidic actuators operated under computer control. An optimization algorithm determines in real time the actuator frequency selected at each iteration based on minimizing the plume centerline temperature. This approach successfully determined an optimal actuator frequency that corresponds to the global minimum for the system without any a priori information about the flow or actuator characteristics. Additional results are presented that address robustness to noise and changes in plant conditions and convergence characteristics.