Michele Gadaleta

Engineering methods and tools enabling reconfigurable and adaptive robotic deburring

According to recent researches, it is desirable to extend Industrial Robots (IR) applicability to strategic fields such as heavy and/or fine deburring of customized parts with complex geometry. In fact, from a conceptual point of view, anthropomorphic manipulators could effectively provide an excellent alternative to dedicated machine tools (lathes, milling machines, etc.), by being both flexible (due to their lay-out) and cost efficient (20-50% cost reduction as compared to traditional CNC machining). Nonetheless, in order to successfully enable high-quality Robotic Deburring (RD), it is necessary to overcome the intrinsic robot limitations (e.g. reduced structural stiffness, backlash, time-consuming process planning/optimization) by means of suitable design strategies and additional engineering tools. Within this context, the purpose of this paper is to present recent advances in design methods and software platforms for RD effective exploitation. Focusing on offline methods for robot programming, two novel approaches are described. On one hand, practical design guidelines (devised via a DOE method) for optimal IR positioning within the robotic workcell are presented. Secondly, a virtual prototyping technique for simulating a class of passively compliant spindles is introduced, which allows for the offline tuning of the RD process parameters (e.g. feed rate and tool compliance). Both approaches are applied in the design of a robotic workcell for high-accuracy deburring of aerospace turbine blades.

Increasing position accuracy and energy efficiency of servo-actuated mechanisms

This paper quantitatively reports about a practical method to improve both position accuracy and energy efficiency of Servo-Actuated Mechanisms (SAMs) for automated machinery. The method, which is readily applicable on existing systems, is based on the “smart programming” of the actuator trajectory, which is optimized in order to lower the electric energy consumption, whenever possible, and to improve position accuracy along those portions of the motion law which are process relevant. Both energy demand and tracking precision are computed by means of a virtual prototype of the system. The optimization problem is tackled via a traditional Sequential-Quadratic-Programming algorithm, that varies the position of a series of virtual points subsequently interpolated by means of cubic splines. The optimal trajectory is then implemented on a physical prototype for validation purposes. Experimental data confirm the practical viability of the proposed methodology.

Determining the Energy-Optimal Base Position of Industrial Robots by Means of the Modelica Multi-Physics Language

A novel technique for determining the energy-optimal base position of common Industrial Robot (IR) is presented. At first, an energy-focused IR model is developed by means of the Modelica/Dymola simulation environment. Then, for a given IR task, a standard but efficient optimization technique is employed, which allows to determine the robot base position corresponding to the minimum energy consumption. A set of graphical maps is finally provided, which allows a rapid estimation of the energy demand along with the time required for the task completion.