INTRODUCTION. Advanced robotic techniques for steel bridge maintenance, Manamperi et al 1

Size: px

Start display at page:

Download "INTRODUCTION. Advanced robotic techniques for steel bridge maintenance, Manamperi et al 1"

Annabel Douglas
5 years ago
Views:

1 Advanced Robotic Technologies for Steel Bridge Maintenance P.B.Manamperi, P.A.Brooks, The Roads and Traffic Authority, New South Wales, Australia, {palitha_manamperi, D.K.Liu, G.Dissanayake, ARC Centre of Excellence for Autonomous Systems, Faculty of Engineering, University of Technology, Sydney, Australia, {dkliu, SYNOPSIS Bridges are essential in transport infrastructure worldwide. Corrosion is the primary cause of failure in steel bridges, and is minimized by painting the steel structure. Steel bridge coating maintenance consists of two procedures: the stripping of rust and deteriorated paint and then repainting the steel; and is one of the biggest expenditure items in bridge maintenance activities. An effective and efficient method of paint stripping is grit blasting, and herein lies the critical problem. Grit blasting is extremely labour intensive and hazardous. Workers have to not only spend long periods of time handling forces of 100N and above, but also need to take precautions to avoid exposure to the dust containing hazardous materials and chemicals. Thus supplementing manual labour in grit blasting with robotic aids will have a significant health, safety and economic impact. In collaboration with the NSW Roads & Traffic Authority (RTA), the ARC Centre of Excellence for Autonomous Systems at the University of Technology, Sydney has been conducting research on a robotic grit blasting system for stripping paint and rust from steel bridges, with the ultimate objective of preventing human exposure to hazardous and dangerous dust containing rust, paint particles, lead and asbestos, relieving human workers from labor intensive tasks, and reducing costs associated with bridge maintenance. A prototype system has been built and is being tested in the laboratory. Advanced robotic techniques for steel structure map building, material type classification, collision avoidance, etc. have been developed. Auxiliary systems such as nozzle force reduction system, safety control systems, etc. have also been designed. This paper will present the robotic system design and the important issues associated with the deployment of an autonomous grit blasting system in the field. INTRODUCTION Bridges facilitate the safe, effective and reliable access and movement of people and goods over streams, roads, railways and other obstacles. They are critical links in the transport network benefiting communities and facilitating the growth of the National and State economies. The RTA takes a strategic approach in managing its assets to ensure that they continue to deliver the required level of service at minimum whole of life cost. The Advanced robotic techniques for steel bridge maintenance, Manamperi et al 1

2 RTA currently maintains about 5,000 bridges in NSW. The RTA bridge stock includes 440 bridges constructed of steel or iron. These bridges were constructed from the early 1860s through to the present day. Steel bridge structures are therefore long term assets. There are two broad classes of steel bridges truss bridges and beam bridges. Bridge steelwork is not designed to be sacrificial during the service life of the structure. Therefore, to deliver design service life, it needs to be protected with coatings to prevent corrosion and metal loss. However, coatings on steel bridges deteriorate over time due to ageing, environmental factors and general wear and tear. They need to be satisfactorily maintained to ensure their effectiveness and design life. Bridges constructed prior to 1900 were constructed from wrought iron. There are 44 wrought iron bridges under the control of RTA. These wrought iron bridges also require maintenance coating. The consequences of corrosion are many and varied. Its effects on the safe, reliable and efficient operations of structures are more serious than the simple loss of a mass of metal. Bridge corrosion prevention is achieved by coating the metal, in order to interpose a corrosion resistant coating between metal and the environment. The action of protective coatings is often more complex than simply providing a barrier between metal and environment. For example, paints may contain a corrosion inhibitor and zinc coating and when applied to iron or steel confers cathodic protection. Until the 1980s, virtually all steel bridges were protected from corrosion by multiple thin coats of lead and chromate-containing alkyd paints applied directly over mill scale on the formed steel. The removal of protective coating systems that contains chromium and lead by grit blasting requires systems to protect the workers as well as the environment. In this research and development project it is intended to develop methodologies to enable a robotic system to perform the grit blasting operation to prepare steel bridges for application of a new paint system. The research and development project carried out by the ARC Centre of Excellence for Autonomous Systems at the University of Technology, Sydney is funded by the RTA and a grant from Australian Research Council. CURRENT INTERNATIONAL PROGRESS There has been increasing research interest in the use of robotics technology in the maintenance of steel structures such as ships, steel bridges, storage tanks, etc. There are commercially available robotic blasters that could be remotely operated for blasting of large vertical or horizontal metal surfaces as commonly found in shipyards and many steel tanks. The working principle of Blastrac equipment is based on the laws of kinetic and mechanical energy. The cleaning operation is performed by steel abrasive (grit or shot) being thrown at high velocity against the steel surface to be cleaned. Advanced robotic techniques for steel bridge maintenance, Manamperi et al 2

A robotic bridge maintenance system [2] (Fig 2) developed for removal of paint and rust from the surfaces of steel beams is more relevant to the current project undertaken for development of a steel

3 The steel grit or shot is thrown by an airless blast wheel rotating at high speed. After the abrasive has impacted the work surface in a high velocity stream, it rebounds back into the machine, taking the rust and old coating with it (Fig1)[1]. A robotic bridge maintenance system [2] (Fig 2) developed for removal of paint and rust from the surfaces of steel beams is more relevant to the current project undertaken for development of a steel bridge maintenance unit. This system consists of a large crane boom, an actuated platform, a four degree of freedom (DOF) robot arm, a gantry table, a vision system (camera) and proximity sensors. This system relies on the availability of CAD drawings of the bridge to minimise the challenges in sensing and surface representation. It is not possible to use this system in a fully enclosed area that is now an environmental requirement for lead paint removal. Fig 1. Blastrac in operation [1] Fig 2. A robotic bridge maintenance system [2] While there are other successful robotic maintenance applications such as crawling robots for cleaning tall buildings or the inside of tanks these robots are limited to the maintenance of large and smooth surfaces and this precludes their use in complex environments encountered in steel bridges. The aim of the project is to initially develop an automated system to remove paint from a girder/beam bridge within a containment. This requires the system to be able to work close to the structure within a containment area and to handle the complexity of the bridge structure and meet the requirements for productivity, safety and cost effectiveness. Advanced robotic techniques for steel bridge maintenance, Manamperi et al 3

4 SYSTEM DESIGN The robotic system for steel bridge maintenance should be able to do the following: Recognise the location of itself within the bridge structure Be able to recognise the surrounding environment Move within the steel elements that need blasting without colliding with steel Be able to grit blast using a blast nozzle and resist the forces due to blasting and be able to carry out the blasting operation efficiently. Do not blast any bridge components other than steel members. Recognise the surfaces so that it could determine the areas to be blasted as well as achievement of required blasting standards. In order to perform the above tasks there are many research and development tasks required. Therefore, in addition to addressing engineering challenges, the system design research is divided into the following research tasks [3]: Environmental awareness Exploring and 3D mapping Blasting surface identification Robot path/motion planning Collision avoidance in an unstructured, new and hazardous environment Selection of a robot Recording the virtual environment and developing the human-robot interface All research outcomes are to be tested and evaluated through a program both in the laboratory and on site. SYSTEM DEVELOPMENT Environment Awareness At the start of a bridge rehabilitation project the robot systems do not have any knowledge of the bridge environment. The knowledge required for the robot to operate in bridge environment could be given by providing CAD drawings of the bridge. However, many old bridges that need paint stripping and painting do not have CAD drawings. Therefore it is necessary to develop a system that will build and record a three dimensional map of the bridge. For this purpose the research team has built a sensor network attached to the robot that could build the geometrical 3D map of the complex structural environment. While building the 3D map of the structure it should avoid potential collision with the unknown and unmapped bridge structure. Exploration and 3D map building Generally the surface geometry of a fixed bridge is static. Therefore it is possible to collect the sensing data relative to the robot s location within the bridge structure.. The data for building a 3D map of the bridge components that needs to be blast cleaned is sensed by the sensor system attached to the robot arm from a known spatial location. The robot collects the sensor data from different robot positions and robot arm poses. These data sets must be transformed to a single global coordinate Advanced robotic techniques for steel bridge maintenance, Manamperi et al 4

frame prior to incorporation in a three dimensional map of the environment (bridge structural components). The robot arm carries the sensor package.

in the coordinate frame of the robot base. Then the data collected from different locations will be combined using advanced simultaneous localization mapping (SLAM) techniques to form a 3D map.

5 frame prior to incorporation in a three dimensional map of the environment (bridge structural components). The robot arm carries the sensor package. As the geometric model of the robot arm is known, in conjunction with joint position describing the arm configuration at the time of data acquisition, the data can be transformed data into 3D points in the coordinate frame of the robot base. Then the data collected from different locations will be combined using advanced simultaneous localization mapping (SLAM) techniques to form a 3D map. In order for the robot to perform the environmental scanning and grit blasting operations, the top end of the robot is attached with a mount designed to accept different equipment attachments including laser scanner, video camera, and gritblasting nozzle apparatus. Figure 3(a) is a picture of an underdeck of a typical steel beam bridge structure; Figure 3(b) shows the to-scale laboratory configuration of the bridge structure. The software simulation model derived from the exploration data is shown in Figure 3(c). In the laboratory setup, the under deck of a steel structure bridge has been chosen as the suitable representation of a typical autonomous grit-blasting environment. The bridge structure is replicated by an exact scale model constructed within the University laboratory. a) Steel bridge structure depicting maintenance environment b) To-scale laboratory bridge section replica c) Simulation environment model constructed from environment exploration data Fig 3. An example bridge structure and the laboratory setup Material Type Classification In a steel bridge structure it is common to have other material types such as timber, concrete and plastic pipes. Therefore it is necessary for the system to clearly identify steel that requires blasting from other materials present in the environment. The research team has developed methods for identifying different material types within the environment using sensors and algorithms to interpret the sensed data. Robot path/motion planning Once the area to be blasted is known, it is essential to plan how to cover the area to be blasted by the robot in the most efficient and effective way. When blasting, the robot has to carry all attachments necessary for blasting (eg. Air hoses, blasting nozzle and grit etc). Therefore it is essential to plan the path of the blast nozzle that gives optimum outputs. It is also necessary to make sure that the developed robot arm path can handle the blasting operation forces without exerting excessive stresses on the robot arm and joints. The project team has developed generic Advanced robotic techniques for steel bridge maintenance, Manamperi et al 5

6 algorithms to determine the best possible consequential path and motion to grit blast the surfaces to achieve predetermined surface finish standard. Simulations and experimental results have demonstrated that use of generic algorithm optimisation procedures have improved the path Collision avoidance in an unstructured, new and hazardous environment The robot needs to move within an unknown environment to carry out the following tasks: Exploration and map building Planning Grit blasting and cleaning Due to the size of the robot relative to the space restricted environment it has to operate in, it is necessary to protect the robot from possible collision within the environment. Therefore collision avoidance is included in all stages of the autonomous operation: (1) Stage 1: Exploration and map building Exploration and map building is the first stage in autonomous operation. At this stage the robot senses and interacts with the environment avoiding potential collisions and collects data to build a 3D map of the bridge structure. To avoid potential collisions within the environment the robot carries a reliable sensor network covering the body of the robot. A control algorithm is developed to control the robot movement for avoiding any potential collisions. (2) Stage 2: Robot path planning At this stage, a 3D map of the environment is constructed using the collected data and algorithms. This 3D map is then used for robot path planning and optimising the robot movement path to ensure consistent and effective grit blasting over the surface and to avoid collision with the structure and to maximise the efficiency of the blasting operation. Collision-free paths are obtained by a path planning algorithm. (3) Stage 3: Grit blasting and cleaning In the grit blasting stage the robot uses the developed robot motion paths and as a result of the prior environmental mapping stage it is not envisaged that any collision would occur. However, it is still necessary to have sensor systems operational to avoid potential collisions due to malfunction of the robot or any sudden change in the environment, i.e. an object left by mistake in the environment. In grit blasting, the blast grit is forced through the blast nozzle at very high velocity which generates a reaction force up to 88 Newtons in the opposite direction of the grit flow. Therefore it is necessary to have a robot with higher payload capacity. However, due to the complex nature of steel structures and the space constraints of the maintenance environment, it is necessary to select a robot that could work within a constrained environment. Advanced robotic techniques for steel bridge maintenance, Manamperi et al 6

7 To address these limitations, a force transfer mechanism is being developed for reduction of the blasting reaction force. Fig 4 shows the mechanism which uses internal pressure of the compressed air from two small delivery hoses to generate a force pushing the nozzle in the opposite direction to the blasting reaction force. The rigidity of delivery hoses deflects the majority of the reaction force through to the hoses to a fixed anchor point. Delivery hose F reaction 88 N (a) (b) (c) Fig 4. Grit-blasting reaction force and a mechanism for force reduction Selection of a robot The selection of the robot for grit blasting was based on the orientation versatility, high speed movements, repeatability and load-handling capability. The 6DOF Denso VM-6083D-W robot has been chosen for the prototype system because it satisfied the required functions. The robot is mounted on a movable platform (Fig 5) giving the prototype system the ability to perform environment exploration and mapping, material type classification and grit-blasting over the majority of the bridge structure to be mapped. Virtual environment and human-robot interface Based on the operational environment the robot has been designed to operate in the following three (3) control modes as detailed below: Manual Mode allowing full operator control. In some complicated locations within the bridge structure only manual mode will be suitable Semi-Autonomous Mode enabling some control by the robot but specifying the defined region where the robot will blast. Autonomous Mode meaning that the robot will be automatically able to commence mapping and blasting in a new uncomplicated part of the bridge with minimal operator involvement. (a) (b) (c) Fig 5. (a) the prototype robotic system; (b) the moving platform and (c) the braking mechanism Advanced robotic techniques for steel bridge maintenance, Manamperi et al 7

8 CONCLUSION AND REMARKS The paper presents progress at the halfway stage of a three year project to develop a robotic system for the grit blast cleaning of steel bridges. While most of the research on technical issues is at a very advanced stage there are many practical issues yet to be solved prior to testing the prototype robot within laboratory conditions. It is anticipated further information regarding the performance of the robotic system will be presented at the Austroads Bridge Conference in May ACKNOWLEDGEMENTS This work is supported by the ARC Linkage Grant (ARC-LP ), by the ARC Centre of Excellence for Autonomous Systems (CAS), the Roads and Traffic Authority, NSW (RTA) and the University of Technology, Sydney. Special thanks go to the research team members of this project for their great efforts and very productive work, to Mr. Paul Robertson for his work on grit-blasting reaction force reduction and hose management system development, and to Mr. James Medlin for his contributions in the moving platform design. DISCLAIMER The opinions expressed in this paper are those of the authors and do not necessarily reflect the policies and practices of the RTA. REFERENCES [1] BLASTMASTER product bulletin November/December 2008 [2] S.J. Lorenc, B.E. Handlon and L.E. Bernold, Development of a robotic bridge maintenance system, Automation in Construction, 9, pp , 2000 [3] D.K. Liu, G. Dissanayake, P.B. Manamperi, P.A. Brooks, et al. (2008), A robotic system for steel bridge maintenance: research challenges and system design, Proceedings of the Australasian Conference on Robotics and Automation (ACRA 08), 3-5 December 2008, Canberra, Australia Advanced robotic techniques for steel bridge maintenance, Manamperi et al 8

A PROTOTYPE CLIMBING ROBOT FOR INSPECTION OF COMPLEX FERROUS STRUCTURES

A PROTOTYPE CLIMBING ROBOT FOR INSPECTION OF COMPLEX FERROUS STRUCTURES G. PETERS, D. PAGANO, D.K. LIU ARC Centre of Excellence for Autonomous Systems, University of Technology, Sydney Australia, POBox