Robotic pipe inspection: system design, locomotion and control
File(s)
Author(s)
Blyth, William Alexander
Type
Thesis
Abstract
The structural integrity of an industrial plant is a key safety and economic consideration for operators
in many sectors including: power generation, oil and gas and petrochemical. To ensure safe operation,
the plant often undergoes scheduled inspection using Non-Destructive Testing (NDT) methods
to detect, size and locate defects such as cracks in welds or corrosion, with the accuracy and repeatability
of inspections being critical to monitoring defect growth. Considering the significant cost of
plant downtime for inspection, there is an economic benefit to be gained by increasing the speed of
inspections, alongside reducing access requirements.
Robotic NDT aims to address some of these issues, with an increase in inspection speed, repeatability
and accuracy, and the added potential to remove human operators from hazardous environments.
Conventional mobile climbing robots are limited by their manoeuvrability, with lateral drift, gravitational
effects and continuous motion being particular issues that remain challenging in the context of
complex geometries.
This thesis aims to investigate the potential for enhanced mobility in mobile climbing robots to precisely
follow inspection paths on cylindrical surfaces in various orientations, without drift. This has
been done through the development of a reduced actuation mecanum wheel platform, using magnetic
adhesion and external position encoding, allowing full translational motion whilst using kinematic
and geometric constraints to prevent rotation. Through dynamic modelling of the platform, a model
based control structure has been synthesised, allowing for improved consistency of performance in
all orientations.
The platform has been applied to representative inspection paths and used to conduct ultrasonic inspection
of test samples, with the capability to correct for lateral drift and cover large inspection areas,
improving the inspection accuracy and area coverage over conventional systems, without requiring
operator intervention.
in many sectors including: power generation, oil and gas and petrochemical. To ensure safe operation,
the plant often undergoes scheduled inspection using Non-Destructive Testing (NDT) methods
to detect, size and locate defects such as cracks in welds or corrosion, with the accuracy and repeatability
of inspections being critical to monitoring defect growth. Considering the significant cost of
plant downtime for inspection, there is an economic benefit to be gained by increasing the speed of
inspections, alongside reducing access requirements.
Robotic NDT aims to address some of these issues, with an increase in inspection speed, repeatability
and accuracy, and the added potential to remove human operators from hazardous environments.
Conventional mobile climbing robots are limited by their manoeuvrability, with lateral drift, gravitational
effects and continuous motion being particular issues that remain challenging in the context of
complex geometries.
This thesis aims to investigate the potential for enhanced mobility in mobile climbing robots to precisely
follow inspection paths on cylindrical surfaces in various orientations, without drift. This has
been done through the development of a reduced actuation mecanum wheel platform, using magnetic
adhesion and external position encoding, allowing full translational motion whilst using kinematic
and geometric constraints to prevent rotation. Through dynamic modelling of the platform, a model
based control structure has been synthesised, allowing for improved consistency of performance in
all orientations.
The platform has been applied to representative inspection paths and used to conduct ultrasonic inspection
of test samples, with the capability to correct for lateral drift and cover large inspection areas,
improving the inspection accuracy and area coverage over conventional systems, without requiring
operator intervention.
Version
Open Access
Date Issued
2017-12
Date Awarded
2018-09
Copyright Statement
Attribution NoDerivatives 4.0 International Licence (CC BY-ND)
Advisor
Rodriguez y Baena, Ferdinando
Cegla, Frederic
Sponsor
Phoenix Inspection Systems Limited
Engineering and Physical Sciences Research Council
Grant Number
EP/I017704/1
Publisher Department
Mechanical Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Engineering Doctorate (EngD)