Analytical wind turbine wake models and wake superposition methods are prevailing tools widely adopted by the wind energy community to predict the power production of wind farms. However, none of the existing wake superposition methods conserve the streamwise momentum. In this study, a novel wake superposition method capable of conserving the total momentum deficit in the streamwise direction is derived theoretically, and its performance is validated with both particle imaging velocimetry measurements and large-eddy simulation results. Detailed inter-method comparisons show that the novel wake superposition method outperforms all the existing methods by delivering an accurate prediction of the power production and the centreline wake velocity deficit, with a typical error of less than 5 % (excluding the near-wake region). Additionally, the momentum-conserving wake superposition method is extended to combine the transverse velocities induced by yawed wind turbines, and the secondary wake steering effect crucial to the power optimization in active wake control is well reproduced.
Plasma synthetic jet actuators (PSJAs), capable of producing high-velocity pulsed jets at high frequency, are well suited for high-Reynolds-number subsonic and supersonic flow control. The effects of energy deposition and actuation frequency on the formation and evolution characteristics of plasma synthetic jets (PSJs) are investigated in detail by high-speed phase-locked particle imaging velocimetry (PIV). Increasing jet intensity with energy deposition is mainly contributed by the increasing peak jet velocity ( $U_{p}$ ), while decreasing jet intensity with actuation frequency is attributed to both the reduced cavity density (primary factor) and the shortened jet duration (secondary factor). The total energy efficiency of the considered PSJA ( $O(0.01\,\%)$ ) reduces monotonically with increasing frequency, while the time-averaged thrust produced by the PSJA is positively proportional to both the deposition energy and the frequency. A simplified theoretical model is derived and reveals a scaling power law between the peak jet velocity and the non-dimensional deposition energy (exponent $1/3$ ). The propagation velocity of the vortex ring attached at the jet front shows a non-monotonic behaviour of initial sharp increase and subsequent mild decay. The peak values for both the propagation velocity and the circulation of the front vortex ring are reached approximately two exit diameters away from the exit. Finally, analysis of the time-averaged flow fields of the issuing PSJ indicates that the axial decay rate of the centreline velocity is proportional to the actuation frequency whereas it is invariant with the energy deposition. The jet spreading rate of the PSJ is found to be higher than steady jets but lower than piezoelectric synthetic jets. Similarly, the entrainment coefficients of the PSJ are found to be twice as high as the values for comparable steady jets.
Machine learning is increasingly used for active flow control. In this experimental study, alternating-current dielectric barrier discharge plasma actuators are deployed for the closed-loop intelligent control of the flow around a cylinder at a Reynolds number of 28 000 based on the velocity feedback from two hot-wire sensors placed in the wake. Variations in the cylinder drag are monitored by a load cell, and the temporal response of the wake flow field is visualized by a high-speed particle image velocimetry system working at 1 kHz. The high-speed control law is operated using a field programmable gate array optimized by genetic programing (GP). The results show that the peak drag reduction achieved by machine learning is of similar magnitude to that of conventional steady actuation (∼15%), while the power saving ratio is 35% higher than with conventional techniques because of the reduced power consumption. Analysis of the best GP control laws shows that the intensity of plasma actuation should be kept at a medium level to maximize the power-saving ratio. When compared with the baseline uncontrolled flow, the best controlled cases constrain the meandering motion of the cylinder wake, resulting in a narrow stabilized velocity deficit zone in the time-averaged sense. According to the results of proper orthogonal decomposition and dynamic mode decomposition, Karman vortex shedding is promoted under the best GP control.
In this study, stereo particle imaging velocimetry measurements are performed at multiple streamwise locations behind a yawed wind turbine to reveal the formation mechanisms of the counter-rotating vortex pair (CVP), and a point vortex transportation (PVT) model is proposed to reproduce the top–down asymmetric kidney-shaped wake (also referred to as a curled wake). Results indicate that the CVP formed behind a yawed wind turbine originates from the complex interactions between the hub vortex and the streamwise components of the blade tip vortices, which is fundamentally different from the case of a yawed drag disk where the hub vortex is absent. Specifically, when the yaw angle exceeds a critical value, a small part of the streamwise vorticity shed from the rotor disk edge switches its sign from negative to positive and subsequently merges with the concentrated hub vortex under mutual induction, creating a patch of positive vorticity; meanwhile, the remaining streamwise vorticity distributed along the rotor edge curls and evolves into another patch of negative vorticity. These two patches of streamwise vorticity essentially constitute the CVP. Based on the physics learnt from the experiments, the non-uniform cross-stream velocity fields are first reconstructed by a cloud of point vortices distributed along the rotor edge and a hub vortex located in the rotor centre, and subsequently used to numerically solve a simplified transportation–diffusion equation of the wake velocity deficit, which altogether constitute the PVT model. This physics-based reduced-order model is the first model capable of accurately reproducing the wake deformation behind a yawed wind turbine.
The plasma synthetic jet actuator (PSJA), also named as sparkjet actuator, is a special type of zero-net mass flux actuator, driven thermodynamically by pulsed arc/spark discharge. Compared to widely investigated mechanical synthetic jet actuators driven by vibrating diaphragms or oscillating pistons, PSJAs exhibit the unique capability of producing high-velocity (>300 m/s) pulsed jets at high frequency (>5 kHz), thus tailored for high-Reynolds-number high-speed flow control in aerospace engineering. This paper reviews the development of PSJA in the last 15 years, covering the major achievements in the actuator working physics (i.e., characterization in quiescent air) as well as flow control applications (i.e., interaction with external crossflow). Based on the extensive non-dimensional laws obtained in characterization studies, it becomes feasible to design an actuator under several performance constraints, based on first-principles. The peak jet velocity produced by this type of actuator scales approximately with the cubic root of the non-dimensional energy deposition, and the scaling factor is determined by the electro-mechanical efficiency of the actuator (O(0.1%–1%)). To boost the electro-mechanical efficiency, the energy losses in the gas heating phase and thermodynamic cycle process should be minimized by careful design of the discharge circuitry as well as the actuator geometry. Moreover, the limit working frequency of the actuator is set by the Helmholtz natural resonance frequency of the actuator cavity, which can be tuned by the cavity volume, exit orifice area and exit nozzle length. In contrast to the fruitful characterization studies, the application studies of PSJAs have progressed relatively slower, not only due to the inherent difficulties of performing advanced numerical simulations/measurements in high-Reynolds-number high-speed flow, but also related to the complexity of designing a reliable discharge circuit that can feed multiple actuators at high repetition rate. Notwithstanding these limitations, results from existing investigations are already sufficient to demonstrate the authority of plasma synthetic jets in shock wave boundary layer interaction control, jet noise mitigation and airfoil trailing-edge flow separation.
Although active flow control based on deep reinforcement learning (DRL) has been demonstrated extensively in numerical environments, practical implementation of real-time DRL control in experiments remains challenging, largely because of the critical time requirement imposed on data acquisition and neural-network computation. In this study, a high-speed field-programmable gate array (FPGA) -based experimental DRL (FeDRL) control framework is developed, capable of achieving a control frequency of 1–10 kHz, two orders higher than that of the existing CPU-based framework (10 Hz). The feasibility of the FeDRL framework is tested in a rather challenging case of supersonic backward-facing step flow at Mach 2, with an array of plasma synthetic jets and a hot-wire acting as the actuator and sensor, respectively. The closed-loop control law is represented by a radial basis function network and optimised by a classical value-based algorithm (i.e. deep Q-network). Results show that, with only ten seconds of training, the agent is able to find a satisfying control law that increases the mixing in the shear layer by 21.2 %. Such a high training efficiency has never been reported in previous experiments (typical time cost: hours).