Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.799457
Title: Autonomous navigation of unknown pipe networks
Author: Brown, Liam
Awarding Body: University of Manchester
Current Institution: University of Manchester
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
Nuclear decommissioning is forecast to cost the UK £121 billion over the next 120 years. This is due to the hazardous environment nuclear material creates making it very expensive to dispose of. In a large number of legacy facilities their contents is unknown and thus must be treated as contaminated waste, unless it is characterised. The aim of this project is to develop an in-pipe robot that can aid in characterisation by autonomously travelling along the unknown pipe networks and recording their geometric data. This enables a map to be generated with the radiation hotspots identified. A key limitation of similar existing systems is the traversal through elbows and junctions. Their primary cornering method is brute force, which can lead to early fatigue and damage to the robots. To allow a robot to safely travel through unknown pipe networks a controlled method for turning is required. The solution to overcome this challenge is the robust, novel wall-pressing robot FURO II, which has been designed and built as part of this project. This 150 mm diameter robot features both active and passive wall-pressing and is able to pass through elbows. It acted as the test platform for the developed control system. The parameters of unknown corners have to determine before a control action can be determined. This is achieved with a novel, low-cost set of feeler sensors and an accompanying algorithm that is able to identify the corner direction with an absolute mean error of 4 degrees, and is the only system to practically estimate the corner radius. These estimated parameters are used as part of an autonomous elbow controller which was tested on FURO II. The controller showed a significant improvement over the brute force method with a reduction in the impulse applied to the robot of 213.97 Ns. A method to visualise the captured data from the robot is proposed and showed to work on the captured data from the previous testing. To summarise, this thesis presents a successful autonomous elbow controller for the navigation of small diameter pipe networks with unknown maps for an in-pipe, wall-pressing, differential drive robot.
Supervisor: Watson, Simon ; Carrasco Gomez, Joaquin Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.799457  DOI: Not available
Keywords: Feeler Sensors ; Pipe Inspection ; Autonomy ; Robotics
Share: