On Increasing the Automation Level of Heavy-Duty Hydraulic Manipulators with Condition Monitoring of the Hydraulic System and Energy-Optimised Redundancy Resolution
Nurmi, Jarmo (2017)
Nurmi, Jarmo
Tampere University of Technology
2017
Rakennetun ympäristön tiedekunta - Faculty of Built Environment
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:ISBN:978-952-15-3976-3
https://urn.fi/URN:ISBN:978-952-15-3976-3
Tiivistelmä
Hydraulic manipulators on mobile machines are predominantly used for excavation and lifting applications at construction sites and for heavy-duty material handling in the forest industry due to their superior power-density and rugged nature. These manipulators are conventionally open-loop controlled by human operators who are sufficiently skilled to operate the machines. However, in the footsteps of pioneering original equipment manufacturers (OEMs) and to keep up with the intensifying demand for innovation, more and more mobile machine OEMs have a major interest in significantly increasing the automation level of their hydraulic manipulators and improving the operation of manipulators. In this thesis, robotic software-based functionalities in the form of modelbased condition monitoring and energy-optimal redundancy resolution which facilitate increased automation level of hydraulic manipulators are proposed.
A condition monitoring system generally consists of software modules and sensors which co-operate harmonically and monitor the hydraulic system’s health in real-time based on an indirect measure of this system’s health. The premise is that when this condition monitoring system recognises that the system’s health has deteriorated past a given threshold (in other words, when a minor fault is detected, such as a slowly increasing internal leakage of the hydraulic cylinder), the condition monitoring module issues an alarm to warn the system operator of the malfunction, and the module could ideally diagnose the fault cause. In addition, when faced with severe faults, such as an external leakage or an abruptly increasing internal leakage in the hydraulic system, an alarm from the condition monitoring system ensures that the machine is quickly halted to prevent any further damage to the machine or its surroundings.
The basic requirement in the design of such a condition monitoring system is to make sure that this system is robust and fault-sensitive. These properties are difficult to achieve in complex mobile hydraulic systems on hydraulic manipulators due to the modelling uncertainties affecting these systems. The modelling uncertainties affecting mobile hydraulic systems are specific compared with many other types of systems and are large because of the hydraulic system complexities, nonlinearities, discontinuities and inherently time-varying parameters. A feasible solution to this modelling uncertainty problem would be to either attenuate the effect of modelling errors on the performance of model-based condition monitoring or to develop improved non-model-based methods with increased fault-sensitivity. In this research work, the former model-based approach is taken. Adaptation of the model residual thresholds based on system operating points and reliable, load-independent system models are proposed as integral parts of the condition monitoring solution to the modelling uncertainty problem. These proposed solutions make the realisation of condition monitoring solutions more difficult on heavy-duty hydraulic manipulators compared with fixed-load manipulators, for example. These solutions are covered in detail in a subset of the research publications appended to this thesis.
There is wide-spread interest from hydraulic manipulator OEMs in increasing the automation level of their hydraulic manipulators. Most often, this interest is related to semi-automation of repetitive work cycles to improve work productivity and operator workload circumstances. This robotic semi-automated approach involves resolving the kinematic redundancy of hydraulic manipulators to obtain motion references for the joint controller to enable desirable closed-loop controlled motions. Because conventional redundancy resolutions are usually sub-optimal at the hydraulic system level, a hydraulic energy-optimised, global redundancy resolution is proposed in this thesis for the first time. Kinematic redundancy is resolved energy optimally from the standpoint of the hydraulic system along a prescribed path for a typical 3-degrees-of-freedom (3-DOF) and 4-DOF hydraulic manipulator. Joint motions are also constrained based on the actuators’ position, velocity and acceleration bounds in hydraulic manipulators in the proposed solution. This kinematic redundancy resolution topic is discussed in the last two research papers. Overall, both designed manipulator features, condition monitoring and energy-optimised redundancy resolution, are believed to be essential for increasing the automation of hydraulic manipulators.
A condition monitoring system generally consists of software modules and sensors which co-operate harmonically and monitor the hydraulic system’s health in real-time based on an indirect measure of this system’s health. The premise is that when this condition monitoring system recognises that the system’s health has deteriorated past a given threshold (in other words, when a minor fault is detected, such as a slowly increasing internal leakage of the hydraulic cylinder), the condition monitoring module issues an alarm to warn the system operator of the malfunction, and the module could ideally diagnose the fault cause. In addition, when faced with severe faults, such as an external leakage or an abruptly increasing internal leakage in the hydraulic system, an alarm from the condition monitoring system ensures that the machine is quickly halted to prevent any further damage to the machine or its surroundings.
The basic requirement in the design of such a condition monitoring system is to make sure that this system is robust and fault-sensitive. These properties are difficult to achieve in complex mobile hydraulic systems on hydraulic manipulators due to the modelling uncertainties affecting these systems. The modelling uncertainties affecting mobile hydraulic systems are specific compared with many other types of systems and are large because of the hydraulic system complexities, nonlinearities, discontinuities and inherently time-varying parameters. A feasible solution to this modelling uncertainty problem would be to either attenuate the effect of modelling errors on the performance of model-based condition monitoring or to develop improved non-model-based methods with increased fault-sensitivity. In this research work, the former model-based approach is taken. Adaptation of the model residual thresholds based on system operating points and reliable, load-independent system models are proposed as integral parts of the condition monitoring solution to the modelling uncertainty problem. These proposed solutions make the realisation of condition monitoring solutions more difficult on heavy-duty hydraulic manipulators compared with fixed-load manipulators, for example. These solutions are covered in detail in a subset of the research publications appended to this thesis.
There is wide-spread interest from hydraulic manipulator OEMs in increasing the automation level of their hydraulic manipulators. Most often, this interest is related to semi-automation of repetitive work cycles to improve work productivity and operator workload circumstances. This robotic semi-automated approach involves resolving the kinematic redundancy of hydraulic manipulators to obtain motion references for the joint controller to enable desirable closed-loop controlled motions. Because conventional redundancy resolutions are usually sub-optimal at the hydraulic system level, a hydraulic energy-optimised, global redundancy resolution is proposed in this thesis for the first time. Kinematic redundancy is resolved energy optimally from the standpoint of the hydraulic system along a prescribed path for a typical 3-degrees-of-freedom (3-DOF) and 4-DOF hydraulic manipulator. Joint motions are also constrained based on the actuators’ position, velocity and acceleration bounds in hydraulic manipulators in the proposed solution. This kinematic redundancy resolution topic is discussed in the last two research papers. Overall, both designed manipulator features, condition monitoring and energy-optimised redundancy resolution, are believed to be essential for increasing the automation of hydraulic manipulators.
Kokoelmat
- Väitöskirjat [4893]