Zhutao Jiang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Zhutao Jiang has a total of 18 co-authors. Here are the top co-authors by number of shared publications:
- Bernd R. Noack
College of Mechatronics and Control Engineering, Shenzhen University , Canghai campus , Shenzhen 518060 , People’s Republic of China
5 shared publications - Guy Yoslan Cornejo Maceda
Chair of Artificial Intelligence and Aerodynamics, School of Mechanical Engineering and Automation, Harbin Institute of Technology 1 , Shenzhen 518055
3 shared publications - Nan Gao
Dalian University of Technology
3 shared publications - Franz Raps
College of Urban Transport and Logistics, Shenzhen Technology University , Shenzhen 518118 , People’s Republic of China
2 shared publications - Tamir Shaqarin
Department of Mechanical Engineering, Tafila Technical University, Tafila 66110, Jordan
2 shared publications - Yiqing Li
University College London
2 shared publications
Latest Publications
- 12 August 2026
- Published by Walter de Gruyter GmbH in Transport Phenomena
- Vol. 1 (3)
Turbulence enhancement with a fan array wind generator using geometrically textured actuation and particle swarm optimization
Fan array wind generators (FAWG) are designed to generate a rich set of turbulent flows reminiscent of those found in natural environments. In this study, we experimentally investigate a square FAWG consisting of 10 × 10 individually controllable fans with 4 cm width and a maximum velocity of 17 m/s. The goal is to maximize the turbulence intensity in the test region. Two approaches for fan operation are investigated: first, geometric texturing of the duty cycle distribution, and second, maximization of the turbulence intensity at selected hot-wire sensors with particle-swarm optimization (PSO). We find that geometric texturing, specifically a checkerboard pattern of active fans, yields a nearly uniform turbulence field driven by jet interactions with turbulence levels around 0.14. However, particle swarm optimization doubles this local turbulence level yielding 0.28 at the chosen sensor locations but sacrifices performance elsewhere. A physical analysis reveals that particle swarm location places the sensors in the shear-layer between active and inactive regions while geometric texturing exploits the large fluctuation at the end of the potential core of the active patches. This region marks the maximum achievable turbulence level. The proposed physical interpretation of texture-based turbulence enhancement is corroborated with a square 40 × 40 fan array wind generator (Guinness World Record 2025) with a width of 3.25 m. Texture-based turbulence control opens a promising avenue to simultaneously command the mean velocity profile and the turbulence level field of a sufficiently large fan array wind generator.
- 12 August 2026
- Published by Walter de Gruyter GmbH in Transport Phenomena
- Vol. 1 (3)
Turbulence enhancement with a fan array wind generator using geometrically textured actuation and particle swarm optimization
Fan array wind generators (FAWG) are designed to generate a rich set of turbulent flows reminiscent of those found in natural environments. In this study, we experimentally investigate a square FAWG consisting of 10 × 10 individually controllable fans with 4 cm width and a maximum velocity of 17 m/s. The goal is to maximize the turbulence intensity in the test region. Two approaches for fan operation are investigated: first, geometric texturing of the duty cycle distribution, and second, maximization of the turbulence intensity at selected hot-wire sensors with particle-swarm optimization (PSO). We find that geometric texturing, specifically a checkerboard pattern of active fans, yields a nearly uniform turbulence field driven by jet interactions with turbulence levels around 0.14. However, particle swarm optimization doubles this local turbulence level yielding 0.28 at the chosen sensor locations but sacrifices performance elsewhere. A physical analysis reveals that particle swarm location places the sensors in the shear-layer between active and inactive regions while geometric texturing exploits the large fluctuation at the end of the potential core of the active patches. This region marks the maximum achievable turbulence level. The proposed physical interpretation of texture-based turbulence enhancement is corroborated with a square 40 × 40 fan array wind generator (Guinness World Record 2025) with a width of 3.25 m. Texture-based turbulence control opens a promising avenue to simultaneously command the mean velocity profile and the turbulence level field of a sufficiently large fan array wind generator.
- 1 June 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (6)
Smart skin separation control with micro-macro machine learning feedback
We introduce a novel distributed control technology sensing and manipulating separation where it is most effective. This control architecture, termed “Smart Skin” in the sequel, comprises an array of control modules and a robust self-learning feedback algorithm. The control modules host an adjustable, fully retractable V-shaped vortex generator (VG), an in-time commandable streamwise jet actuator and pressure sensors. The self-learning feedback algorithm is based on a physics-informed distributed-input distributed-output structure. Each control module listens to local (micro) and global (macro) sensing signals and command the VG adaptively and the jet actuators in time. These physics-informed feedback laws are optimized with gradient-enriched Machine Learning Control (MLC) and termed micro-macro (m2) MLC in the sequel. The Smart Skin separation control is experimentally demonstrated for the backward-facing ramp with 30 multi-modal actuators and 42 pressure sensors at a Reynolds number of 32 000 based on height. m2MLC achieves a 125% increase in base pressure compared to the unforced case. Particle image velocimetry measurements show a fully attached flow by actuation. We expect numerous applications of smart-skin separation control on smooth surfaces, for instance, on helicopter and air taxi nacelles, on wings near the stall conditions, and inside diffusers.
- 1 June 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (6)
Smart skin separation control with micro-macro machine learning feedback
We introduce a novel distributed control technology sensing and manipulating separation where it is most effective. This control architecture, termed “Smart Skin” in the sequel, comprises an array of control modules and a robust self-learning feedback algorithm. The control modules host an adjustable, fully retractable V-shaped vortex generator (VG), an in-time commandable streamwise jet actuator and pressure sensors. The self-learning feedback algorithm is based on a physics-informed distributed-input distributed-output structure. Each control module listens to local (micro) and global (macro) sensing signals and command the VG adaptively and the jet actuators in time. These physics-informed feedback laws are optimized with gradient-enriched Machine Learning Control (MLC) and termed micro-macro (m2) MLC in the sequel. The Smart Skin separation control is experimentally demonstrated for the backward-facing ramp with 30 multi-modal actuators and 42 pressure sensors at a Reynolds number of 32 000 based on height. m2MLC achieves a 125% increase in base pressure compared to the unforced case. Particle image velocimetry measurements show a fully attached flow by actuation. We expect numerous applications of smart-skin separation control on smooth surfaces, for instance, on helicopter and air taxi nacelles, on wings near the stall conditions, and inside diffusers.
- 1 September 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (9)
Jet mixing optimization using a flexible nozzle, distributed actuators, and machine learning
In this paper, we introduce the first jet nozzle allowing simultaneous shape variation and distributed active control, termed “Smart Nozzle” in the sequel. Our Smart Nozzle manipulates the jet with an adjustable flexible shape via 12 equidistant stepper motors and 12 equidistantly placed inward-pointing minijets. The mixing performance is evaluated with a 7 × 7 array of Pitot tubes at the end of the potential core. The experimental investigation is carried out in three steps. First, we perform an aerodynamic characterization of the unforced round jet flow. Second, we investigate the mixing performance under five representative nozzle geometries, including round, elliptical, triangular, squared, and hexagonal shapes. The greatest mixing area is achieved with the square shape. Third, the symmetric forcing parameters are optimized for each specified nozzle shape with a machine learning algorithm. The best mixing enhancement for a symmetric active control is obtained by the squared shape, which results in a 1.93-fold mixing area increase as compared to the unforced case. Symmetrically unconstrained forcing achieves a nearly 4.5-fold mixing area increase. The Smart Nozzle demonstrates the feasibility of novel flow control techniques that combine shape variation and active control, leveraging the capabilities of machine learning optimization algorithms.
- 1 September 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (9)
Jet mixing optimization using a flexible nozzle, distributed actuators, and machine learning
In this paper, we introduce the first jet nozzle allowing simultaneous shape variation and distributed active control, termed “Smart Nozzle” in the sequel. Our Smart Nozzle manipulates the jet with an adjustable flexible shape via 12 equidistant stepper motors and 12 equidistantly placed inward-pointing minijets. The mixing performance is evaluated with a 7 × 7 array of Pitot tubes at the end of the potential core. The experimental investigation is carried out in three steps. First, we perform an aerodynamic characterization of the unforced round jet flow. Second, we investigate the mixing performance under five representative nozzle geometries, including round, elliptical, triangular, squared, and hexagonal shapes. The greatest mixing area is achieved with the square shape. Third, the symmetric forcing parameters are optimized for each specified nozzle shape with a machine learning algorithm. The best mixing enhancement for a symmetric active control is obtained by the squared shape, which results in a 1.93-fold mixing area increase as compared to the unforced case. Symmetrically unconstrained forcing achieves a nearly 4.5-fold mixing area increase. The Smart Nozzle demonstrates the feasibility of novel flow control techniques that combine shape variation and active control, leveraging the capabilities of machine learning optimization algorithms.
- Open Access
- 14 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 495-498
Fast Self-Learning of Turbulence Feedback Laws Using Gradient-Enriched Machine Learning Control
We apply genetic programming (GP) to solve the most complex turbulence control problems. We simultaneously optimize up to dozens of parameters and dozens of control laws, listening to dozens of sensor signals with 100 to 1000 short test runs. Unlike reinforcement learning, our implementation of GP does not require any meta parameter tuning and is many orders of magnitudes faster than vanilla versions thanks to numerous enablers: 1. Parameter optimization over many dozens of experiments and simulations. 2. Smart formulation of control laws by preprocessing inputs and outputs. 3. Gradient-enriched simplex optimization of promising subspace. This work focuses on the resulting algorithm, referred to as gradient-enriched Machine Learning Control (gMLC). The applications comprise learning a multiple-input multiple-output control law to stabilize the flow past a cluster of three rotating cylinders in numerical simulations and the stabilization of the shear layer in an open cavity flow experiment. The learning acceleration achieved by gMLC opens the path to complex and time-limited experiments, including evaluation in varying operating conditions and optimization of distributed-input distributed-output control laws, i.e., functions with O(100) inputs and outputs.
- 14 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 495-498
Open AccessFast Self-Learning of Turbulence Feedback Laws Using Gradient-Enriched Machine Learning Control
We apply genetic programming (GP) to solve the most complex turbulence control problems. We simultaneously optimize up to dozens of parameters and dozens of control laws, listening to dozens of sensor signals with 100 to 1000 short test runs. Unlike reinforcement learning, our implementation of GP does not require any meta parameter tuning and is many orders of magnitudes faster than vanilla versions thanks to numerous enablers: 1. Parameter optimization over many dozens of experiments and simulations. 2. Smart formulation of control laws by preprocessing inputs and outputs. 3. Gradient-enriched simplex optimization of promising subspace. This work focuses on the resulting algorithm, referred to as gradient-enriched Machine Learning Control (gMLC). The applications comprise learning a multiple-input multiple-output control law to stabilize the flow past a cluster of three rotating cylinders in numerical simulations and the stabilization of the shear layer in an open cavity flow experiment. The learning acceleration achieved by gMLC opens the path to complex and time-limited experiments, including evaluation in varying operating conditions and optimization of distributed-input distributed-output control laws, i.e., functions with O(100) inputs and outputs.
- 1 January 2024
- Published by American Institute of Aeronautics and Astronautics (AIAA) in AIAA Journal
- Vol. 62 (1) , 291-301
Aerodynamic Characterization of a Fan-Array Wind Generator
Experimental assessment of safe and precise flight control algorithms for unmanned aerial vehicles under gusty wind conditions requires the capability to generate a large range of velocity profiles. In this study, we employ a small fan-array wind generator that can generate flows with large spatial and temporal variabilities. We perform a thorough aerodynamic characterization, operating the fans uniformly from low to maximum levels. Particle image velocimetry and hot-wire measurements indicate a jetlike flow with a nearly uniform core, which monotonously contracts in the streamwise direction and the surrounding, growing unsteady shear layers. These complex dynamics result in a limited region with a desired flow profile and turbulence level. The experimental results shed light on the flow generated by a full-scale fan-array wind generator, and they indicate the need for further improvements via properly designed add-ons and dedicated control algorithms.
- 1 January 2024
- Published by American Institute of Aeronautics and Astronautics (AIAA) in AIAA Journal
- Vol. 62 (1) , 291-301
Aerodynamic Characterization of a Fan-Array Wind Generator
Experimental assessment of safe and precise flight control algorithms for unmanned aerial vehicles under gusty wind conditions requires the capability to generate a large range of velocity profiles. In this study, we employ a small fan-array wind generator that can generate flows with large spatial and temporal variabilities. We perform a thorough aerodynamic characterization, operating the fans uniformly from low to maximum levels. Particle image velocimetry and hot-wire measurements indicate a jetlike flow with a nearly uniform core, which monotonously contracts in the streamwise direction and the surrounding, growing unsteady shear layers. These complex dynamics result in a limited region with a desired flow profile and turbulence level. The experimental results shed light on the flow generated by a full-scale fan-array wind generator, and they indicate the need for further improvements via properly designed add-ons and dedicated control algorithms.