Indoor environments are continuously subject to the effects of thermal radiation: short-wave radiation that brings in solar heat, and long-wave radiation that acts to redistribute heat. However, the effects of the latter are often overlooked. To fully understand and predict the effects of long-wave radiation indoors, spatial variations in indoor temperature must be considered. This paper develops an analytical framework to describe coupled radiative and convective heat transfers within simple representations of thermally stratified rooms, a minimal description requiring only five thermally active bodies. The model, termed ‘AR5B’, predicts the coupled heat fluxes from each thermal body, enabling the effects of long-wave radiation to be simply mimicked within ventilation flow models. Data from six full-scale experiments show that predictions from AR5B are of suitable accuracy. Moreover, when the AR5B results are then used to mimic the effects of radiation within a simplified ventilation flow model, the resulting thermal stratification is reasonably predicted. We show this to be in stark contrast to comparable cases but without accounting for the effects of long-wave radiation – this finding underscores that long-wave radiation plays a significant role in determining indoor environments and should be incorporated more routinely in predictive models.

We develop a data-driven approach to infer aerodynamic drag directly from canopy geometry by learning interpretable low-dimensional latent representations of canopy configurations, without resolving the flow field. A PixelCNN-based variational autoencoder with latent disentanglement regularisation and auxiliary observable regression is used to identify latent factors that encode the geometric features most relevant to drag. Latent traversals and mutual information analysis are employed to quantify the physical relevance of individual latent dimensions and to assess how modelling choices affect information retention. Applied to laboratory measurements of heterogeneous canopy arrays, the learned latent space organises canopy configurations according to their aerodynamic impact and enables accurate drag prediction. The results demonstrate that physically informed latent modelling provides a compact and interpretable pathway for linking complex canopy geometry to aerodynamic drag.

We present the design, construction and initial experimental validation of the Northwestern Polytechnical University Taylor–Couette (NPU-TC) apparatus, specifically developed to explore turbulent Taylor–Couette flows under conditions relevant to ultra-high-speed rotating machinery. The apparatus features an inner cylinder capable of rotating at speed of up to 10 000 rpm, corresponding to a Taylor number $Ta = 6.4 \times 10^8$, with an exceptionally narrow annular gap of 2.8 mm, yielding a radius ratio ($\eta$) of 0.98. Axial-scanning particle image velocimetry is employed here for the first time in air-based TC flows at such extreme conditions, which enables detailed velocity measurements without intrusive disturbances. Our velocity measurements demonstrate the absence of large-scale coherent flow structures, indicating a transition into the ultimate turbulence regime characterised by very thin boundary layers and nearly uniform velocity distributions in the bulk region. The NPU–TC apparatus thus represents a significant advance in experimental capabilities, providing critical insights into turbulent flow behaviour in high-speed rotating machinery.

This study examines the control capabilities of an array of spanwise-invariant roughness strips applied on a swept-wing boundary layer (BL) dominated by a cross-flow instability (CFI) that is forced by periodically spaced discrete roughness elements to a monochromatic wavelength. Several configurations of strip arrays are investigated, varying their height, width and chordwise periodicity. Infrared thermography is employed to track the impact on the BL transition location. Optimal configurations are identified, extending laminar flow by up to 10 % of the wing chord. Additionally, BL forced by patches of randomised surface roughness are considered, better representing realistic wing surfaces. In this scenario, the application of strip arrays with optimal geometry extends the laminar portion of the BL by almost 10 % chord and beyond when combined with a discrete roughness element array. Time-averaged particle image velocimetry (PIV) velocity fields are acquired to monitor the CFI amplitude for the various configurations. The BL spectral content in the spanwise direction is used to characterise the chordwise behaviour of individual disturbance modes, whose amplitude is found to be reduced by up to 17 % for the optimal strip configuration.

Power loss mechanisms in large wind farms are complex due to the multiscale nature of wind farm aerodynamics. Recent studies based on ‘two-scale momentum theory’ have brought new insights into this field; however, most of them have been limited to idealised wind farm scenarios. To better understand power loss mechanisms in real wind farms, in this study, we extend the framework of the two-scale momentum theory to non-ideal turbine design and layout scenarios, and then introduce simple analytical models to account for the associated power losses. This extension provides a holistic view of how turbine design, layout, operating conditions and atmospheric conditions collectively determine the amounts of different types of power losses in real wind farms, including the losses due to turbine-wake interference (i.e. ‘internal’ power loss) and farm-atmosphere interaction (i.e. ‘external’ power loss). We also present a simple iterative method for calculating the optimal farm induction factor that maximises the overall farm power for a given set of conditions, including the atmospheric boundary layer height. Analogously to blade-element momentum theory playing a key role in wind turbine design optimisation, the present theory may play a key role in wind farm design optimisation.

We present and analyse observational data from a highly instrumented classroom computer laboratory and develop a multi-zone model to describe its mechanical ventilation and mixing regime. The laboratory houses 70 workstations that are used heterogeneously in time and space, in a manner similar to a generic office environment. Our model predicts CO$_2$ concentration in the laboratory, accounting for air exchange between the occupied classroom and its ceiling plenum, and by parametrising irreversible mixing in each zone. Applying the model to our measurements helps identify critical components in the ventilation network, as highlighted by a strong separation of the time scales characterising the flow response. On the one hand, this time scale separation leads to a simplified model describing the CO$_2$ transport. On the other hand, it suggests that the forced exchange of volume between the room and the plenum is ‘overdriven’ in that reduced energy operation could be achieved without compromising air quality. More generally, our modelling approach offers a systematic method to enhance energy efficient ventilation of multi-zone systems.

Longitudinal vortices produced by a swirl-mixing grid are experimentally explored in an upscaled model of nuclear fuel assembly. The flow is mapped using particle image velocimetry in several planes downstream of the grid. The flow, an isothermal flow geometrically similar to that in one of the standard nuclear reactors, is compared between basic grids, swirl grids and the case without fuel rods, allowing for a link to previous studies of longitudinal vortex lattices. Individual vortices are recognised using a custom-made algorithm. Analysis of vortices shows that the meandering is enhanced by the presence of fuel rods and by the presence of an upstream swirl grid. The vortex core radii do not grow in the constrained case. There is a weak anticorrelation between the vortex velocity and the actual meandering amplitude. The neighbouring vortices show a weak correlation in their circumferential velocities or energies, but they do not display any significant correlations of positions or meandering amplitudes, cutting down any hypothetical “vortex dancing”.

Chirped coherent Rayleigh–Brillouin scattering (CRBS) is a flow diagnostic technique that offers high signal-to-noise ratios and nanosecond temporal resolution. To extract information of dilute gas flow, experimental spectra must be compared with theoretical predictions derived from the Boltzmann equation. In this work, we develop a MATLAB code that deterministically solves the Boltzmann equation (with a modelled collision kernel for the inverse power-law potential) to compute CRBS spectra, enabling each line shape to be obtained in approximately one minute. We find that the CRBS spectrum is highly sensitive to the intermolecular potential and that rapid chirping generates fine ripples around the Rayleigh peak along with spectral asymmetries.

Wind energy is a sustainable and plentiful form of clean energy. The vertical axis wind turbine (VAWT) is one type of cost-effective, acoustically quieter and lightweight turbines. The two mainstream types of VAWTs – Darrieus (lift type) and Savonius (drag type) – have contradictory strengths and weaknesses. Darrieus turbines possesses high efficiency but suffer from poor self-starting, while Savonius turbines start easily with poor aerodynamic performance. In this study, a novel VAWT with adaptive Darrieus–Savonius hybrid blades has been designed. Wind-tunnel experiments assessed the effectiveness of the proposed design and compared it against a conventional Darrieus-type rotor with similar dimensions. The results showed that the static torque coefficient was improved by over 65% and the self-starting wind speed was reduced from 8 to 6 ms−1. The adaptive blades can remain in the Savonius configuration at low rotation speed, facilitating self-starting, and automatically transition to the Darrieus configuration at higher rotation speed, integrating and leveraging the strengths of the two types of VAWTs.

The surface pressure distribution over a circular cylinder with a small, full-span, triangular bump is examined. The geometry of the bump is an isosceles triangle, the height of which is varied from 1.33 % to 5.33 % of the diameter of the cylinder and positioned between $60^{\circ }$ and $120^{\circ }$. The Reynolds number ($Re = V_{\infty}D/\nu$, where $V_\infty$ is the velocity of the freestream, $D$ is the diameter of the cylinder and $\nu$ is the kinematic viscosity) is varied between $1.1 \times 10^5$ and $1.8 \times 10^5$. The lift and drag are estimated through the surface integral of pressure over the cylinder. The results show that the smallest bump acts as a trip for the lower Re and orientations before $70^{\circ }$, leading to a separation farther upstream than in the case of no bump. For larger bumps, Re and orientation angles, the bump acts as a spoiler and fully separates the boundary layer at the bump. In addition, the surface pressure upstream of the bump is strongly dependent on the bump position. The lift is highest for bump position less than $90^{\circ }$ and decreases significantly with increasing bump location angle. The drag is less sensitive to the position of the bump. These findings have implications for predicting the forces on bluff bodies due to small asymmetric surface geometry features and extension to applications such as atmospheric flow over topography.

This study quantifies the viscous interaction between propeller tip vortices and a turbulent boundary layer developing over a semi-elliptic leading-edge plate, located downstream. The experimental wind-tunnel set-up is designed to be representative of the tractor–propeller–wing configuration. Using stereoscopic particle image velocimetry and static wall-pressure measurements, the near-wall flow topology is resolved over the plate, semi-immersed in the propeller slipstream. The results show that the interaction exhibits high spatio-temporal coherence and is dominated by a coupling between primary and secondary vortical structures. Two distinct interaction regions are identified relative to the tip-vortex core: on the inboard side, towards the slipstream interior, the boundary-layer flow experiences strong velocity gradient transitions and amplified near-wall vorticity. The flow on the outboard side, moving out of the slipstream, exhibits wall-parallel velocity deficits and vorticity lift-up consistent with unsteady vortex-induced separation mechanisms. Spanwise velocity induced by the wall-normal component of the primary vortex connects these two regions, with the secondary vortex structure identified as enhancing boundary-layer lift-up on the outboard side. Although no local flow reversal occurs under the tested conditions, localised shear amplification and vorticity roll-up indicative of separation-like behaviour were observed. These findings advance the understanding of viscous slipstream–boundary-layer interaction and its implications for tractor–propeller–wing integration.

Fan-array wind tunnels are an emerging technology to prescribe wind fields through grids of individually controllable fans. This design is especially suited for the turbulent, dynamic, non-uniform flow conditions found close to the ground, and has enabled applications from entomology to flight on Mars. However, due to the high dimensionality of fan-array actuation and the complexity of unsteady flow, the physics of fan arrays are not fully characterised, making it difficult to prescribe arbitrary flow fields. Accessing the full capability of fan arrays requires resolving the map from time-varying grids of fan speeds to three-dimensional unsteady flow fields, which remains an open problem. In this paper, we study the case of constant fan speeds and time-averaged streamwise velocities with one homogeneous spanwise axis. We produce a proof-of-concept surrogate model by fitting a regularised linear map to a dataset of fan-array measurements. We use this model as basis for an open-loop control scheme to prescribe flow profiles subject to constraints on fan speeds. We experimentally validate our model and control scheme, provide a physical interpretation of our model as a superposition of self-similar jet profiles and conclude that the physics relating constant fan-array speeds to time-averaged streamwise velocities are dominantly linear.

Accurate and efficient modelling of cardiac blood flow is crucial for advancing data-driven tools in cardiovascular research and clinical applications. Recently, the accuracy and availability of computational fluid dynamics methodologies for simulating intraventricular flow have increased. However, these methods remain complex and computationally costly. This study presents a reduced-order model (ROM) based on higher-order dynamic mode decomposition (HODMD). The proposed approach enables accurate reconstruction and long-term prediction of left ventricle flow fields. The method is tested on two idealized ventricular geometries exhibiting distinct flow regimes to assess its robustness under different hemodynamic conditions. By leveraging a small number of training snapshots and focusing on the dominant periodic components representing the physics of the system, the HODMD-based model accurately reconstructs the flow field over entire cardiac cycles and provides reliable long-term predictions beyond the training window. The reconstruction and prediction errors remain below 5 % for the first geometry and below 10 % for the second, even when using as few as the first three cycles of simulated data, representing the transitory regime. Additionally, the approach reduces computational costs with a speed-up factor of at least $10^{5}$ compared with full-order simulations, enabling fast surrogate modelling of complex cardiac flows. These results highlight the potential of spectrally constrained HODMD as a robust and interpretable ROM for simulating intraventricular hemodynamics. This approach shows promise for integration in real-time analysis and patient specific models.

The urban canopy affects wind in complex ways, making it challenging to predict wind-driven natural ventilation and cooling in buildings. Using large eddy simulations of coupled outdoor and indoor airflow, we study how the surrounding urban canopy and wind angle influence ventilation rates through four ventilation configurations: cross, corner, dual-room and single-sided. Flow visualisations demonstrate how both large-scale flow patterns and local interference effects can influence ventilation rates by 50 %–85 %. In general, lower density canopies give higher ventilation rates, and wind angles that align with a direct path between two openings also lead to higher ventilation rates. However, interference effects from surrounding buildings can significantly change the local wind speed and direction, thus also changing ventilation rates. The magnitude of these interference effects depends on both the wind angle and surrounding building geometries. The effect of wind angle is less pronounced in a higher density canopy, where the urban canopy geometry more strongly guides the flow. The results demonstrate that the canopy’s effect on ventilation rates is much more complex than those suggested by existing natural ventilation parametrisations.