Investigation on the Effects of Vertical Camera Placement on the Accuracy of Optical Plume Velocimetry

Timothy J Crone developed a non-invasive flow measurement technique known as optical plume velocimetry to acquire flow rates using video image analysis. In estimating the flow rate, a video camera was used to capture image sequence containing two-dimensional projections of the three-dimensional flow field. An important parameter in using a video camera is the camera placement. The effects of vertical camera placement on the accuracy of optical plume velocimetry has not been studied and is the motivation for this research. A change in vertical position of a camera changes the view of images captured by a video camera, hence resulting in a possibility for data collection error. The main motivation of this research is to assess the accuracy of optical plume velocimetry technique in estimating flow rate at various vertical camera placements. The accuracy of optical plume velocimetry is assessed with seven different vertical camera angles of -30˚, -20˚, -10˚, 0˚, 10˚, 20˚ and 30˚, and temporal cross correlation algorithm technique was used to find the time delay. Three simulations were carried out for each camera placement to compare the mean flow rate with the actual flow rate. The percentage error obtained for each case of vertical camera placement was very high. Despite that, the percentage error was found to be highly consistent when the simulations were repeated. In that case, the camera angle of 0˚ was chosen to be the datum and a constant was introduced to correct the estimated flow rate. Referring to 0˚, the best estimate of flow rate was obtained when the camera is placed at an angle of -20˚ and it was also evident that camera placements at negative angles yielded better results compared to camera placements at positive angles. In conclusion, angle of -20˚ yielded the best results when the datum was taken to be 0˚, with a percentage error of 5.51%. Also, the negative angle camera placements yielded better results compared to the positive angle camera placements, with the maximum percentage error of 13.29% at angle of -30˚.