This file was created by the Typo3 extension sevenpack version 0.7.16 --- Timezone: CEST Creation date: 2023-06-03 Creation time: 15-35-07 --- Number of references 4 article Noack_25082020 Journal of Fluid Mechanics 2020 897 27 46 An artificial intelligence (AI) control system is developed to maximize the mixingrate of a turbulent jet. This system comprises of six independently operated unsteadyminijet actuators, two hot-wire sensors placed in the jet and genetic programmingfor the unsupervised learning of a near-optimal control law. The ansatz of thislaw includes multi-frequency open-loop forcing, sensor feedback and nonlinearcombinations thereof. Mixing performance is quantified by the decay rate of thecentreline mean velocity of the jet. Intriguingly, the learning process of AI controldiscovers the classical forcings, i.e. axisymmetric, helical and flapping achievable fromconventional control techniques, one by one in the order of increased performance,and finally converges to a hitherto unexplored forcing. Careful examination of thecontrol landscape unveils typical control laws, generated in the learning process, andtheir evolutions. The best AI forcing produces a complex turbulent flow structurethat is characterized by periodically generated mushroom structures, helical motionand an oscillating jet column, all enhancing the mixing rate and vastly outperformingothers. Being never reported before, this flow structure is examined in variousaspects, including the velocity spectra, mean and fluctuating velocity fields and theirdownstream evolution, and flow visualization images in three orthogonal planes, allcompared with other classical flow structures. Along with the knowledge of theminijet-produced flow and its effect on the initial condition of the main jet, theseaspects cast valuable insight into the physics behind the highly effective mixing ofthis newly found flow structure. The results point to the great potential of AI inconquering the vast opportunity space of control laws for many actuators and sensorsand in optimizing turbulence.� mixing enhancement, jets, turbulence control https://www.cambridge.org/core/journals/journal-of-fluid-mechanics/article/artificial-intelligence-control-of-a-turbulent-jet/FD1BA8BD9F20C797DC6FFBAE0173187B 0022-1120 https://doi.org/10.1017/jfm.2020.392 Y.Zhou D.Fan B.Zhang R.Li B. R.Noack article Noack12062020 Experiments in Fluids 2020 61 152 13 A fast triple-parameter extremum seeking method is applied for jet control based on the pioneering work of Gelbert et al.(J Process Control 22(4):700, 2012). The simultaneous adaptation of three input parameters takes less time than the singleinputadaptation of each parameter combined. The key enablers are phase-shifted sinusoids for the input each of which isevaluated by an extended Kalman filter (EKF). An acceleration of the adaption is obtained by a combined EKF couplingthe output to all inputs. The method is illustrated for an analytical optimization problem and experimentally demonstratedfor a turbulent jet mixing control. The considered Reynolds numbers ReD based on the jet exit diameter and velocity are5700, 8000 and 13,300. The main jet is manipulated by a pulsed radially injected minijet which is varied by a mass flowcontroller and an electromagnetic valve up to high frequencies. The mixing performance is characterized by the centerlinejet decay rate and monitored by a hot-wire sensor five diameters downstream at the end of the potential core. The proposedtriple-parameter extremum seeking method optimizes the actuation mass flow ratio, frequency and duty cycle. The decayrate increases 11-fold from the unforced reference value of 0.05 to the optimal actuation level of 0.56. The reproducibilityis demonstrated with various initial actuation parameters. Moreover, the adaptive control robustly tracks the optimal openloopactuation for varying ReD ; the optimal decay rate remains unchanged given the mass flow rate, frequency and dutycycle are optimized. The unforced and actuated flow are investigated with hot wires and visualizations. The three-input ESsignificantly outperforms a two-parameter optimization for the same configuration in multiple respects (Wu et al. in AIAAJ 56(4):1463, 2018): First, the jet decay rate is 8% faster. Second, the convergence time for three parameters is only 25% ofthe adaptation period of two parameters when ReD is varied. Finally, the current steady-state error is 45% less than that ofthe two-parameter optimization. We expect the proposed triple-parameter extremum seeking to be applicable for a largerange of flow control experiments.� https://link.springer.com/article/10.1007/s00348-020-02953-3 14321114,07234864 https://doi.org/10.1007/s00348-020-02953-3 D.Fan Y.Zhou B. R.Noack article Noack21122020 Physics of Fluids 2020 32 125117 1-18 A machine learning control (MLC) is proposed based on the explorative gradient method (EGM) for the optimization and sensitivity analysis of actuation parameters. This technique is applied to reduce the drag of a square-back Ahmed body at a Reynolds number Re = 1.7 × 105. The MLC system consists of pulsed blowing along the periphery of the base, 25 pressure taps distributed on the vertical base of the body, and an EGM controller for unsupervised searching for the best control law. The parameter search space contains the excitation frequency fe, duty cycle α, and flow rate blowing coefficient Cm. It is demonstrated that the MLC may cut short the searching process significantly, requiring only about 100 test runs and achieving 13% base pressure recovery with a drag reduction of 11%. Extensive flow measurements are performed with and without control to understand the underlying flow physics. The converged control law achieves fluidic boat tailing and, meanwhile, eliminates the wake bistability. Such simultaneous achievements have never been reported before. A machine-learned response model is proposed to link the control parameters with the cost function. A sensitivity analysis based on this model unveils that the control performance is sensitive to fe and α but less so to Cm. The result suggests that a small sacrifice on performance will give a huge return on actuation power saving, which may provide important guidance on future drag reduction studies as well as engineering applications. https://aip.scitation.org/doi/full/10.1063/5.0033156 1070-6631 https://doi.org/10.1063/5.0033156 B.Zhang Y.Zhou Y.Fan B. R.Noack article Campos-Delgado2003 Transactions on Control Systems Technology 2003 11 4 429-447 D. U.Campos-Delgado B.Schuermans K.Zhou C. O.Paschereit E.Gallestey A.Poncet