This documentation addresses the modelling, design and control of a mine-grade electro-hydraulic manipulator. The purpose of the study is to develop and demonstrate automatic control tools that can be adopted to hydraulically powered multi-link mining equipment operating in underground mines to improve their safety and productivity performance.
The challenge to such automation is achieving the desired positioning precision required by the nature of the mining function being performed. The obstacles include low manufacturing accuracy, highly nonlinear dynamic characteristics of the fluid power systems, lack of suitable sensors ruggedised for mining environment, and lack of proven automatic control techniques that can cope with the unpredictable and unstructured nature of underground mining environment.
This documentation starts with a close look at the underground mining operation itself and the variety of tasks currently carried out electro-hydraulically manipulated multilink equipment operated by humans. The generic challenges facing robotics and automation in underground metalliferous mines are described based on the experience of the author in three major underground metalliferous mines of Australia. Particular attention is given to the analysis of the rock bolting operation.
An electro-hydraulic manipulator is designed and constructed to simulate the most pertinent characteristics of the manipulators found in current underground mining machinery, with the aim of developing and demonstrating on this manipulator adaptive control techniques, which afterwards can be readily implemented on real machinery operating underground. Imperfections of real machinery, such as loose joints and associated backlash, inadequate structural stiffness, and non-linear behaviour of the hydraulic actuators, are built into the experimental manipulator to make it a realistic representation of mining manipulators.
A computer simulation of this manipulator is developed using Matlab/Simulink. The most critical system parameters are recognised through such simulation and the simulation also serves an initial proving ground for the control alternatives being considered. The numerical values of some of these parameters are determined through preliminary system identification tests and others are determined on-line during operation.
A self-tuning adaptive control scheme is implemented for this manipulator. The performance of this scheme is compared against a fixed-gain PID controller. It is seen that the fixed-gain controller is not capable of satisfying the control design criteria in the face of variable conditions without human operator intervention. The adaptive control scheme has a superior performance and fulfils all criteria set for the control operation.