Particle diffusometry (PD) is a technique of measuring the diffusion coefficient of a fluid sample by seeding it with tracer particles and observing their motion under a microscope. In microfluidic set-ups, the observed particles are often defocused and their motion is affected by factors such as fluid flow, which leads to high errors for conventional and deep learning-based PD (DPD) algorithms. This work improves the performance of DPD models by updating their architecture, avoiding temporal averaging in the input, and exploring the impact of various choices during training. These models provide state-of-the-art performance for generalised datasets regardless of particle shapes, concentration, flow or image noise and are called DPD-v2. These models provide a mean absolute error of 0.09$\unicode{x03BC}$m2s−1 for Gaussian particles and 0.07$\unicode{x03BC}$m2s−1 for defocused particles, which is 2x–4x lower errors as compared with the two following best methods. The performance of DPD-v2 models increases with crop size and the use of multiple stacks of images. The outputs of the DPD-v2 models were compared against the outputs from conventional algorithms on Gaussianised experimental no flow datasets, which provided < 0.5$\unicode{x03BC}$m2s−1 mean absolute difference. Hence, the DPD-v2 models can be used in real-world scenarios.

The stability of supercritical Couette flow has been studied extensively, but few measurements of the velocity field of flow have been made. Particle image velocimetry (PIV) was used to measure the axial and radial velocities in a meridional plane for non-wavy and wavy Taylor–Couette flow in the annulus between a rotating inner cylinder and a fixed outer cylinder with fixed end conditions. The experimental results for the Taylor vortex flow indicate that as the inner cylinder Reynolds number increases, the vortices become stronger and the outflow between pairs of vortices becomes increasingly jet-like. Wavy vortex flow is characterized by azimuthally wavy deformation of the vortices both axially and radially. The axial motion of the vortex centres decreases monotonically with increasing Reynolds number, but the radial motion of the vortex centres has a maximum at a moderate Reynolds number above that required for transition. Significant transfer of fluid between neighbouring vortices occurs in a cyclic fashion at certain points along an azimuthal wave, so that while one vortex grows in size, the two adjacent vortices become smaller, and vice versa. At other points in the azimuthal wave, there is an azimuthally local net axial flow in which fluid winds around the vortices with a sense corresponding to the axial deformation of the wavy vortex tube. These measurements also confirm that the shift-and-reflect symmetry used in computational studies of wavy vortex flow is a valid approach.