PIV – Particle Image Velocimetry
Aerodynamics: Lesson 18
PIV – Particle Image Velocimetry.
Particle Image Velocimetry (abbreviated as PIV) is an optical method of measuring the motion field of a fluid. It provides, in a specific section of the flow, the projection of the field of the instantaneous velocity vector on the section itself.
The flow is inseminated with tracer particles (“seeding“) with density as close as possible to that of the fluid to be studied, in order to follow its motion as closely as possible. At this point the section to be examined lights up with two consecutive and close pulses of laser light, converted by means of an optical apparatus into blades of light. The particles refract the light, which is captured by a camera with the help of a synchronizer. This gives two images of the position of the particles at two different instants, close to each other. Comparing the two images we obtain the field of the vector displacement of the particles on the plane of the light blade. Assuming that the seeding has been correctly chosen, the particles will have followed the motion of the fluid, so dividing the displacement by the time interval Δt that elapses between the detection of the first and second images, the flow velocity field is obtained, which will be as close to the instantaneous speed as the Δt is small.
The main problems concern the seeding concentration, the non-two-dimensionality of the motion field and the choice of the time interval.
If the concentration is low, in fact, although it is easy to follow the particles between the two images, an exhaustive description of the motion field is not obtained: there will be too few points where the speed is known. If, on the other hand, the concentration is high, it is difficult to identify the particles in the images, even with the aid of software. With an average concentration, in any case, it is not possible to follow the particles with the naked eye, so to obtain the displacement the images are divided into smaller windows, called “interrogation windows“, as many as the points where you want know the speed, and it is processed with autocorrelation or cross-correlation algorithms.
If there is a non-negligible component of the speed in an orthogonal direction to the laser plane, the PIV does not record this component and, in addition to parallax problems that distort the recorded speed, some of the particles, in the time elapsing between the two images, they could escape from the thin area illuminated by the laser.
The choice of the time interval is decisive in this sense: if it is too long, the probability that the particles exit the laser and are therefore not taken in the second image is great. Furthermore, even in the case of two-dimensional flow, if the interval is too long, there is a loss of information between the two images and the speed obtained is an average speed which can no longer be assimilated to the instantaneous speed. On the other hand, if the interval is excessively short, the displacement is too small and the disturbances due to the “noise” and the imperfect correlation between the pairs of windows becomes predominant with respect to the actual displacement itself, as an effect, there is a speed range which does not correspond to the real one. The Δt value is in the order of ∼10 µs.
PIV, SPIV and Tome PIV.
There are also methods derived directly from the PIV that allow you to measure all three components of the speed field. The main ones are stereoscopic PIV (or stereo PIV or SPIV) and tomographic PIV (or tome PIV).
The PIV stereo always operates on a predetermined plane which is that of the blade of light, like the simple PIV. To reconstruct the flow velocity vector on the plane in a three-dimensional way, this technique uses the same principle used by our eyes to perceive the depth at which the objects are located: we take two pairs of images with two cameras spaced and / or inclined to each other thus having two projections of a three-component vector on two distinct planes, that is, the image planes of the two cameras. By applying geometric reconstruction algorithms, we go back to three-dimensional speed. The main problems, in addition to those of simple PIV, arise for reasons of perspective and concern the processing of the images captured by the cameras to be able to obtain images on which the geometric reconstruction can be applied.
The PIV tome, on the contrary, operates on a volume and measures the three components of the velocity vector in the whole volume considered. This technique is based on the same principle of simple PIV, however enlarged to 3 dimensions. The entire volume is illuminated with two short-delay laser pulses and the movement of the particles is measured with any number of cameras. In this case, great difficulties arise in uniquely locating the particles, since, not knowing their position in an orthogonal direction to the image plane of each camera, the projections of the numerous particles can overlap. A typical consequence is the generation of “ghost particles”, non-existent in reality but recreated by the software for the reconstruction of the particles precisely because of this.