PIV implementation in flow control : design and real-time considerations : traineeship report

The purpose of this report is to give an assessment of the available techniques to estimate the velocity vector field in a fluid using images of the fluid and, possibly, tracers in the fluid. The selected technique in this report is the particle image velocimetry (PIV) because has been shown as a suitable technique to obtain good estimates of the velocity field of a flow in different scenarios. As a more specific objective, it is desired to develop a systematic procedure to design a PIV sensor and guarantee the achievement of the control goals. The images to be used will resemble the pair 1.1 and 1.2, or the pair 5.2 and 5.3. The first pair of images were captured in a real setup with a long exposure time allowing the tracers motion to be registered in the image. The second pair correspond to synthetic images generated using simulations of a flow. There are many techniques for motion estimation using image processing such the ones presented in [1, 3, 5, 17, 25, 29, 37] among others, most of these techniques are designed or adjusted for a specific motion estimation problem and will be briefly commented in chapter 2. The Particle Image Velocimetry (PIV) technique will be studied in more detail in chapter 3 as the technique chosen to address our motion problem because it is one of the most used techniques for fluid motion estimation. Some authors have been interested on implement the PIV technique in real-time such as [4, 7, 15, 23, 32, 33, 34, 40], the results and proposals made in these works will be discussed in chapter 3.1, meanwhile chapter 6 will present the considerations that must be done in order to obtain a suitable PIV sensor