Ein Gesamtkonzept zur Optimierung der Positioniergenauigkeit von Parallelstrukturen
- An overall concept to optimize the positioning accuracy of parallel robots
Wahle, Martin; Corves, Burkhard (Thesis advisor)
Dissertation / PhD Thesis
Aachen, Techn. Hochsch., Diss., 2014
The positioning accuracy of parallel robots is a decisive factor for evaluating process quality and stability. Within this doctoral thesis an overall concept for assessing and optimizing the positioning accuracy of parallel robots is presented. The developed methods are validated with an existing parallel robot with five degrees of freedom. Yet, due to the generic approach the methods can be adapted to any further non-redundant parallel architecture. Basically, the introduced method distinguishes between positioning errors which are accessible as control errors within the control system and those errors which can only be evaluated by redundant sensors or external measurement equipment. The internal control errors are minimized by utilizing a model-based feedforward control. The developed model considers inertia as well as friction effects. Within the next step, a new control architecture is introduced. Within this control approach the eigenfrequencies and damping ratios of the end-effector modes can be defined separately and independently for each degree of freedom. To map the properties of the entire control loop within the simulation properly the transfer behaviors of the relevant mechatronic components are modeled and identified with measurements. It is shown that by applying the introduced coupled control architecture the stability ranges for one PD-controlled separate actuator match the stability range for the entire parallel robot. Moreover, the stability reduction due to elasticities of mechanic components like the mobile platform and the universal joints are evaluated with multi-body-simulation models. Based on this, a parameter setting method is presented which factors in the measured stability range and modifies the system’s gain margins with respect to each separate degree of freedom such that an optimal control parameter setting is found. To reduce the positioning errors which are not accessible within the control a structural stiffness model of the system is introduced. This model is able to quantify the static motion errors of the end-effector due to process and inertia forces. Furthermore, the model can evaluate the elasticity contribution of each mechanical component. With the help of this model the mechanical rotation restriction is identified to be one major elasticity contributor. As a result, a newly developed rotation restriction is presented. The last part of the thesis deals with the reduction of kinematic errors due to tolerances and assembly imperfections. To achieve this, an available calibration method is improved and applied to the parallel robot. Starting with the optimization of suitable measurement poses the identification is performed with the help of an external measurement system. By utilizing this method the static positioning accuracy can be improved by more than 80 per cent.
- Chair and Institute of Mechanism Theory, Machine Dynamics and Robotics