Experimental study on centerline velocities of a rectangular capture hood under realistic conditions

Capture hoods are an important component of a local ventilation system designed to reduce exposures to airborne contaminants. The velocity at any point along the centerline of the hood (V_x) is currently estimated using one of many predictive equations developed since the 1930s. It is unproven that those predictive equations for V_x are accurate, despite the prodigious number of studies concerning them. Among other issues, almost all experimental verifications were conducted for conditions that were either unrealistically ideal without competing air currents (e.g., zero cross draft) or were not described. This study measured values of V_x along the midline using Particle Image Velocimetry (PIV) at distances of 1-14 inches in front of a rectangular capture hood. The experiments were conducted in a large wind tunnel (9' × 12' × 40', H × W × L) using a heated, breathing, anthropomorphically sized manikin. Three 0 degree draft velocities (V_draft = 4, 14, and 50 ft/min) were tested, all directed toward the hood face and the back of the manikin (if present). For each value of V_draft, the velocity fields were measured in a factorial design with and without the manikin, and with and without a worktable underneath the hood. An ideal condition was represented by a freestanding hood at the 4 fpm draft. Nonideal conditions included the presence of a worktable or manikin, and the combination of table and manikin. Each condition was tested at the three levels of V_draft. The experimental results found significant effects (p < 0.001) for V_draft, the presence of the manikin, the presence of the worktable, and all combinations of those factors. The effects of the independent variables were most pronounced at distances greater than 10 in (25.4 cm) from the hood face. It is concluded that none of the previously published models accurately predicted V_x under the realistic conditions tested in this study. A satisfactory model will have to include terms for V_draft and the presence of a worktable and a worker.