School of Mechanical Engineering Seminar
Wednesday, Nov 14, 2012 at 15:00
Wolfson Building of Mechanical Engineering, Room 206

Criteria for the Resolution of Human Arm Redundancy Problem

Barak Kashi

MSc Student of Prof. Moshe Fuchs, School of Mechanical Engineering, Tel-Aviv University, and Dr. Moshe Brand, Dept. of Mechanical Engineering and Mechatronics, Ariel University Center of Sameria.

The human arm is kinematically redundant in a point task, in which the tip of the index finger is brought to a preselected point in 3D space. Therefore, there is theoretically an infinite number of postures that accomplish the task. Despite the redundancy of human arm posture, several previous experiments showed that the final posture of the human arm in a point task is invariant. In this work a posture prediction model of a four degree of freedom arm was formulated. The model combines two criteria, in order to study prediction capabilities of multi-criteria models compared to single criterion models.  The two criteria that comprise the combined model are: Minimal Angular Displacement (MAD), which minimizes the length of the movement path in joint space, and maximal Joint Range Availability (JRA), which finds the posture that is farthest away from joint limits. To validate the model's predictions, a 3D pointing experiment was carried out. Four subjects performed point to point movements from and to eight targets that were distributed in space. For every movement that was collected in the experiment, three final postures were calculated according to MAD, JRA and the combined model. The calculated postures and the measured postures were linearly correlated. It was concluded that combining two criteria into one posture prediction model with a 50/50 weight ratio improves final posture predictions (slope 1.1, r²=0.68), compared to single criterion models (slope 0.71, r²=0.48 for MAD model, and slope 0.97, r²=0.55 for JRA model). Redundancy resolution of articulated serial mechanisms such as the human or robotic arm has applications in the fields of wearable robotics, computer graphics and ergonomics.

School of Mechanical Engineering Seminar
Wednesday, November 14, 2012 at 15:00
Wolfson Building of Mechanical Engineering, Room 206

 Two Phase Liquid-Liquid Flow in Pipe Bends

Yona Kaplan

MSc Student of Prof. Amos Ullmann and Prof. Neima Brauner

The centrifugal force acting on a fluid in a pipe bend is proportional to the density of the fluid. It is thus reasonable to assume that in two phase flow the heavy phase would be forced to the outer wall of the pipe at the bend. However, this is not always the case. Depending on the physical properties of the fluids, for certain ranges of liquid and gas flow rates, the light phase is forced to the outer wall. This phenomenon was experimentally observed in studies concerning gas-liquid flow in helical tubes and named 'film inversion'.

In the current study the phenomena was investigated in liquid-liquid flow with very low density difference between the liquids (less than 1%). The experimental test section consists of 2 mm inside diameter glass pipe. The tested flow rates of the two liquid phases correspond to laminar flow. The flow upstream the bend is co-current down-flow, where in most cases the flow pattern was core- annular flow.  The downstream section of the bend is horizontal. The flow pattern through bends was studied experimentally by photographing the flow through pipe bends at varies flow rates. In addition the system was modeled using 'Fluent' software. The inlet boundary condition was derived from analytical velocity profile in fully-developed annular flow in the appropriate flow rates.

Generally, good agreement was obtained between the simulation and the experimental observation of the flow phenomena trough the bend. A criterion for 'film inversion' is suggested and tested.