470 D.C. Reid et al. 2 The Longwall Mining Process Longwall mining is a full extraction mining process in which large panels of a coal seam up to 5m thick are completely mined. An indicative longwall panel is 250m wide by 2000m long. A longwall mining system used in this process is comprised of three main components: a shearer, an armoured face conveyor (AFC) and a roof support system. A longwall shearer as shown in Figure 1, is up to 15 metres long, weighs 90 tonnes and typically extracts a one metre slice of the coal seam as it travels back and forth across the panel along rails attached to the AFC. Portions of the roof support system and AFC can also be seen in Figure 1. The roof support system can have over 200 individual hydraulic support modules which collectively provide temporary support of the roof material above the extracted coal seam. The load capacity of each support can exceed 1000 tonnes. As the shearer moves across the coal seam, large hydraulic push rams attached to the roof support modules are used to progressively advance the AFC, and thereby the shearer rails, behind the shearer in a snake-like manner. A shuffle is required at each end of travel to advance the end portions of the AFC. The snake and shuffle are indicated in Figure 2. As the longwall equipment progresses in this manner, the roof material collapses into the void left behind the advancing system. The complete longwall system is a mobile semi-autonomous underground mining machine weighing in excess of 700 tonnes with each of the three main components operating under largely independent and proprietary control systems. Fig.1. Alongwall shearer showing the leading and trailing drum. Aportion of the roof support canopyand AFC are also visible. Shearer Guidance: AMajor Advance in Longwall Mining 471 Fig.2. Schematic of shearer path and AFC profile in plan viewfor atypical cutting mode. The advancing AFC (solid line) bends to form asnake. The shearer path (dashed line) shuffles at each end of travel 2.1 The Need forAutomation Automation of the longwall mining process has always held the lure of increased productivity butmore recently is being drivenbyissues of occupational health and safety.The presence of hazardous gases, respirable dust and the inherent danger of personnel working in close proximity to large mobile mining equipment is becoming increasingly unacceptable. There have been manyattempts worldwide overanumber of decades to achieve full automation of the longwall mining process [4]. Equipment manufacturers have invested heavily in ongoing development of their respective proprietary control systems and yet, to date, personnel are still required to routinely work in hazardous production areas and to manually control the mining process. Previous automation attempts have in large part been stymied by the inability to accurately determine the three-dimensional path of the longwall shearer as it systematically progresses through the coal panel. Without this information there is no absolute reference for controlling the motion of the equipment and reliable, sustained automation can not be achieved. Automated face alignment is amajor deliverable of the Landmark project. Face alignment refers to the process of maintaining adesired path for the shearer in the horizontal plane as it “slices” across the coal face. In order to minimize mechanical stresses on the mining equipment and maximize production, the face is generally required to be straight and at ageodetic heading nominally orthogonal to the direction of panel progression. Face alignment is presently achievedbymanually aligning the position of the AFC and each roof support with reference to astring line deployed across the face for the purpose. This adjustment is typically required about every eight hours of operation and is both time consuming and non-productive. 3Automated Face Alignment Automated face alignment is achievedusing the horizontal position information from the shearer-mounted inertial navigation system (INS) as it travels along the 472 D.C. Reid et al. AFC. As represented in Figure 2, at any particular time, the AFC at the i th support is moved from the n th to the n + 1 th shear cycle. The distance d is computed from INS information which is gathered during the n 1− th pass so as to achieve a desired face profile. In control theory terms, the desired face profile, which includes the absolute geodetic heading, is the system set point. The desired face profile is typically a straight line but other non-straight profiles could be advantageous under certain geological conditions. The control system output is the proportional control of the AFC movement via the roof support system. Negative feedback is provided by the shearer-mounted INS which measures the three-dimensional position of the shearer at closely sampled points across the face. Position error in the AFC proportional control is represented as a system disturbance. Due to INS processing requirements, shearer position data is batch processed at the end of each full face traverse so that the profile corrections made during the n + 1 th shear cycle are computed from data gathered throughout the n th cycle. Profile correction values are calculated as shown in Figure 3. The correction values (solid arrows) at positions corresponding to each roof support module are normalized to be zero at points where no correction is required (point D) and negative valued elsewhere. For each increment of panel progression the required advance distance at each roof support module is then computed by the roof support control system as the addition of the correction value and a constant default advance distance (typically 1m). A correction of zero at all points across the face will result in the longwall progressing the default distance. This strategy ensures that the mining process can continue under open loop control during periods where the correction information is unavailable. An example of the shearer path under open-loop control as measured by the shearer mounted INS is shown in Figure 4. The vertical path projected onto the vertical plane correctly follows the natural undulations in the coal seam and is consistent across the multiple shear cycles. The horizontal component as projected Fig.3. Diagrammatic representation of the relationship between the desired face profile (dashed line AorC), actual face profile (solid line B), normalised position correction values (solid arrows), required advance distance (dashed arrows) and the resulting face profile (solid line F). Fig.4. Three-dimensional path of the moving shearer throughout anumber of shear cycles as measured by the INS. onto the horizontal plane highlights the departure in the face profile from the desired straight line due to accumulated position errors in the open-loop face alignment control system. It is interesting to note that at the time this data wascollected, the longwall operators determined by visual inspection that the face profile wasstraight. In the automated face alignment system these position errors are minimized by systematically adjusting the AFC movement at each roof support module. Full underground trials of the automated face alignment system are planned for second half of 2003 at Beltana Colliery,NSW Australia. The performance of the automated face alignment system is critically dependent on the accuracyand precision of the INS. 3.1 Stabalised Inertial Navigation System Inertial navigation systems are subject to position drift with time mainly as aresult of the numerical double integration required to compute three-dimensional position Shearer Guidance: AMajor Advance in Longwall Mining 473 474 D.C. Reid et al. from three axis acceleration. Dead-reckoning techniques using external odometry can be used to improve short term position stability but systematic drift can still occur if the incremental motion of the INS is not exactly along the measured geodetic heading. High performance INS, such as the military grade units used in this project, typically use GPS aiding to correct this inherent drift. This integrated approach combines the short term accuracy of the INS with the long term stability of GPS. In the underground mining application GPS is not available and so other bias correction strategies were developed. Without effective bias correction the INS derived shearer path may diverge (or converge) in both the horizontal and vertical components [5]. An example of this divergence is apparent in Figure 5 which represents the uncorrected data of Figure 4. Fig.5. Three-dimensional path of the moving shearer throughout anumber of shear cycles as measured by the INS without bias correction. Divergence in the shearer path due to this bias is apparent. Shearer Guidance: AMajor Advance in Longwall Mining 475 INS stabilisation techniques generally rely on externally available position or ve- locity information such as GPS, vehicle odometry or zero velocity updates (ZUPTs). In the Landmark project INS stability has been achievedbyrecognising the (al- most) closed-path of shearer travelthroughout each shear cycle. In normal mining operations the horizontal closing distance for each cycle is either fixed or can be independently determined. This information is used in the automated face alignment system to back-correct the shearer path at the completion of each shear cycle. Simi- larly,back-correction in the vertical plane can be achievedbased on independently surveyed levels which are generally available at the panel boundaries. 4Equipment Interconnection Standard The Landmark automation strategy combines newenabling technologies with exist- ing proprietary control systems from the major international equipment manufactur- ers. These manufacturers are working closely with the Landmark project to integrate their proprietary control systems while maintaining market differentiation and pro- tecting proprietary knowledge. The practical success of the Landmark automation project therefore depended heavily on establishing an industry acceptable data and control standard across the various equipment components. This standard needed to: • Take advantage of the existing Ethernet cabling and network infrastructure available in manymines • Allowmine operators to mix and match mining equipment from various vendors • Be non-proprietary and easily maintained • Support future development and system expansion. ALandmark specification has nowbeen developed and accepted by the industry for each of the major equipment components. This specification is based on the newly developed EtherNet/IP control and information protocol managed and promoted by the Open DeviceNet Vendor Association (ODVA ). EtherNet/IP combines the provenand popular application layer protocol of DeviceNet and ControlNet with the convenience, bandwidth and flexibility of Ethernet hardware and internet protocols. The choice of EtherNet/IP givesthe mining industry the ability to leverage the rapid advances being made in Ethernet technology drivenbythe vast enterprise market and increasingly by the industrial control market. This ability wasdemonstrated in the Landmark project by using inexpensive commercial off-the-shelf wireless Ethernet hardware to provide arelatively high bandwidth data link to the moving shearer.The link wasestablished using anumber of wireless access points distributed at fixed locations across the longwall face and aworkgroup bridge installed on the shearer.These units required very little modification for the underground environment and featured channel diversity and hand-offmechanisms for increased link reliability.These units also offered the convenience of web-based remote administration and configuration. 476 D.C. Reid et al. 5 Summary and Conclusions Longwall mining accounts for a large portion of underground coal production world- wide. The industry is seeking ways to improve productivity and safety for mining personnel. Significant advances in longwall automation are being achieved through the industry sponsored Landmark project. A major deliverable of this project is automated face alignment which promises productivity and safety benefits to the industry. INS-based techniques are being successfully employed in this project to accurately measure the three-dimensional path of the longwall shearer. INS provides the enabling technology for automated face alignment that is paving the way towards full automation of the longwall mining process. Techniques have been developed to ensure the long-term position stability of the INS. A specification for the in- terconnection of underground mining equipment has been published as part of the Landmark project. This specification is based on the newly developed EtherNet/IP control and information protocol and well positions the mining industry to benefit from rapid advances in network and industrial control technology. References 1. D. C. Reid, D. W. Hainsworth and R. J. McPhee, “Lateral Guidance of Highwall Mining Machinery Using Inertial Navigation”, 4th International Symposium on Mine Mechani- sation and Automation,, pp B6-1 B6-10, Brisbane, Australia, 1997. 2. http://www.longwallautomation.org 3. D. W. Hainsworth and D. C. Reid (2000), Mining Machine and Method, Australian Patent PQ7131, April 26, 2000 and US Patent, May 12, 2000. 4. A. L. Craven and I. R. Muirhead, “Horizon Control Technology for Selective Mining in Underground Coal Mines”, The Canadian Mining and Metallurgical Bulletin, Vo lume 93, Number 1040, May, 2000. 5. D. C. Reid, D. W. Hainsworth, J. C. Ralston, and R. J. McPhee, “Longwall Shearer Guid- ance using Inertial Navigation”, Australian Coal Association Research Project C9015 report, June 2001. Development of an Autonomous Conveyor-Bolting Machine forthe Underground Coal Mining Industry Jonathon C. Ralston, Chad O. Hargrave,and David W. Hainsworth Mining Automation CSIROExploration and Mining Technology Court, Pullenvale, Q4069, Australia. jonathon.ralston@csiro.au Abstract. This paper describes the development of anew autonomous conveyor and bolting machine (ACBM) used for the rapid development of roadways in underground coal mines. The ACBM is amobile platform fitted with four independent bolting rigs, bolt storage and delivery carousel, coal receiving hopper and through-conveyor for coal transport. The ACBM is designed to operate in concert with astandard continuous mining machine during the roadway development process to automatically insert roof and wall bolts for securing the roadway.This innovative machine offers significant benefits for increasing personnel safety and improving productivity.The paper describes the core sensing and processing technologies involved in realizing the levelofautomation required by the ACBM, which includes online roof monitoring, roadway profiling, navigation, and automatic control of drilling and bolting processes. 1Introduction The CSIROMining Automation is agroup that concentrates on developing and applying modern automation technology to mining equipment and systems. Au- tomation technology has significant potential to meet the mining industry’songoing need to improve productivity and safety.This is achievedbydeveloping newma- chines and mining processes, creating predictive maintenance and hazard monitoring systems, adding intelligent sensing and processing systems to existing equipment, and by removing personnel from hazardous environments. One of the keyareas for automation in underground coal mining is the devel- opment of the core roadway infrastructure. Roadway development is acomplex, expensive and time-consuming process using acombination of different mining machinery to cut alattice network. The main performance bottleneck in roadway development is the need to constantly halt mining to allowthe installation of sup- porting bolts to prevent the roadway from collapsing. Moreover, the current practice of manually drilling and bolting is one of the most dangerous tasks in underground coal mining, involving significant safety concerns for mine personnel. Areal need therefore exists for arapid roadway development system to minimize personnel exposure to hazardous areas of unsupported roof, as well as to improve the overall production rate of this vital mining activity. S. Yuta et al. (Eds.): Field and Service Robotics, STAR 24, pp. 477–486, 2006. © Springer-Verlag Berlin Heidelberg 2006 478 J.C. Ralston, C.O. Hargrave, and D.W. Hainsworth In an effort aimed at addressing this roadway development problem, a new mining machine, known as the Autonomous Conveyor-Bolting Machine (ACBM), has been designed. The ACBM is a mobile platform fitted with independent bolting rigs, coal receiving hopper, bolt storage and delivery system, and a through-conveyor for coal transport. It is designed to follow the path of a continuous miner as it drives a new roadway, automatically inserting roof and wall bolts. Figure 1 shows the ACBM during factory testing. Figure 2 shows the placement of machinery associated the rapid roadway development process, with a leading continuous miner, the ACBM and a shuttle car for coal transport. Fig.1. The ACBM showing tramming platform, automatic bolting rigs, receiving hopper,bolt storage and delivery system. Fig.2. Placement of machinery associated the rapid roadway development process, with a leading continuous miner,the ACBM and ashuttle car for coal transport. The ACBM showing tramming platform, automatic bolting rigs, receiving hopper,bolt storage and delivery system. Development of an Autonomous Conveyor-Bolting Machine 479 2ACBM Functional Overview The ACBM is amobile platform fitted with independent bolting rigs, coal receiving hopper,bolt storage and delivery system, and athrough-conveyor.Itisdesigned to followthe path of acontinuous miner as it drivesanewroadway,automatically inserting roof and wall bolts. While creating the roadway,coal cut from the tunnel is also transferred via athrough-conveyor belt which leads to the shuttle-car which transports the coal to the surface. The cycle time for placement of arow of four bolts is approximately fiveminutes, allowing amachine advancement rate of approxi- mately 15 m/hour with a1.2-metre rowspacing. The ACBM uses acombination of processing systems in order to provide online roof monitoring, roadway profiling, navigation, and control of drilling and bolting processes. The ACBM control system thus has twofundamental operating modes, namely tramming and bolting. 2.1 ACBM Processing and Control As the ACBM may be interposed between various production and coal haulage machines, the system has been designed to work either independently or in concert with amodified remote controlled miner with remote controlled bolting capabilities. The ACBM uses acentralized unit for the intelligent control and coordination of a set of distributed computing and sensing modules. The central unit is responsible for ACBM tramming, bolting, and conveyor tasks, as well as generic supervisory tasks such as link/device integrity monitoring and system-wide safety.The aim of the control system is to execute the necessary control overthe robotic bolting and motion systems in order to implement the required bolting pattern. The bolting control system is designed to place up to six bolts in arow,oriented from avertical placement to an outward angle of 15 degrees with amaximum vertical reach of 3.7m from the floor.The block diagram in Figure 3shows the control hierarchy between the central control unit and associated signal processing components to achieve this goal. Although the ACBM is designed with fully automatic drilling and bolting capabilities, the system can be set into semi-automatic or manual modes. This permits the operator to elect the operational mode. Agraphical user interface allows operators to interact with the system. 2.2 SoftwareArchitecture The integrated signal processing component technologies are implemented at three orthogonal layers: Validation, execution, and functional. The validation layer has the highest priority and is responsible for top-levelintersystem and inter-machine coordination, system mode resolution, integrity monitoring and other safety related logic decisions. The execution layer controls and coordinates the dynamic execution of main functions such as drilling and bolting sequencing, profiling and drill mon- itoring. The functional layer contains modules that encapsulate all device specific interface and control details (such as drivers and communication protocols) for the sensors and actuators of the machine. [...]... is a well-kept safety research mine and is thus in complete contrast with the abandoned Florence mine Bruceton is well-ventilated, and the floors are dry, groomed, and clear of debris The walls and ceiling are coated with concrete to maintain the integrity of the mine, and there are no low-hanging obstructions Yet, Bruceton and Florence are both coal mines and thus share similar size, shape and general... Abandoned Mines 493 Fig 5 Left: The forward scanning laser rangefinder and custom tilt unit Right: Low-light camera and infrared LED ring 5.2 Conditions Mathies had been used as a coal-haulage route, effectively a tunnel, that ran for more than 1km through a mountain until it closed in early 2000 The corridor was 5-7 meters wide with a drainage ditch on the right-hand side and was expected to be dry and. .. unpredictable terrain of abandoned mine corridors, Groundhog provides a means of conclusively establishing the extents of an abandoned mine by direct observation 2 Chassis and Electronics Groundhog’s chassis is based on the union of two front halves (steering columns and differentials) from recreational all-terrain vehicles, allowing all four of its wheels S Yuta et al (Eds.): Field and Service Robotics, STAR... 1999 2 J C Ralston and D W Hainsworth, “An Autonomous Bolting Machine for Rapid Roadway Development,” IEEE Int Conf Mech and Mach Vision in Practice (M2VIP ’2001), Hong Kong, August 2001 3 C Ye and J Borenstein, “Characterization of a 2-D Laser Scanner for Mobile Robot Obstacle Negotiation,” Proc Inter Conf Robotics and Automation, Washington, pp 251 2-2 518, May 2002 4 J C Ralston and D W Hainsworth,... implementation is clean, efficient, cross platform and readily scalable One of the key benefits of the framework is that it allows for rapid prototyping and development in an environment suitable for deploying the ACBM processing and control algorithms The scripts are in plain-text format and of arbitrary length At program run time, the scripts are read, verified and evaluated and thus no source code recompilation... submerged Fig 2 shows the robot at its farthest point, sunk in 4 0-5 0 cm of sludge and water 3.4 Results For these teleoperated experiments the data from the laser scanners was timestamped and logged for post-processing using scan matching techniques as described in [5] The two-dimensional map shown in Fig 3 shows the entrance in the lower-left corner and represents approximately 45 meters from the entrance... behaviour codification and run-time validation The VPLC language employs constructs that would be typically expected of modern automation-oriented languages such as locally and globally scoped variables, timers, temporal and persistent data objects, fast IO access, conditional evaluations, assignment operators, transitions and state definitions The VPLC engine is based entirely on the C++ standard template... techniques[3] such as ground penetrating radar and seismic surveying that rely on geophysical models and assumptions to infer the existence of underground voids The accuracy and reliability of these systems, however, are highly dependent on the depth of the mine and the structure and composition of the surrounding strata Robots, on the other hand, may operate safely and reliably inside mines independent of... its locomotive load Equipped with six deep-cycle lead-acid batteries providing more than 6 kWh of energy, Groundhog has a locomotive range greater than 3 km Groundhog is just under 1 meter tall and is 1.2 meters wide, originally intended to navigate the breach between the Quecreek and Saxman mine, which was 1.2 meters tall and 2 meters wide The investigation and subsequent sealing of the breach prevented... of Abandoned Mines 489 a hydraulic locomotive system and a 225kg steel enclosure for the majority of its electronics The explosion-proof enclosure houses an industrial DC motor that drives the hydraulic system A 300 MHz PC/104+ CPU and associated I/O electronics also occupy the enclosure along with a hydraulic manifold with six solenoid actuators All outgoing power lines are computer-controlled and . the shearer at closely sampled points across the face. Position error in the AFC proportional control is represented as a system disturbance. Due to INS processing requirements, shearer position. ACBM. The computing systems also need to operate in presence of potentially explosive gases and thus need to be housed in aspecial manner.Thisrequires that the ruggedised processing modules be placed. manymines • Allowmine operators to mix and match mining equipment from various vendors • Be non-proprietary and easily maintained • Support future development and system expansion. ALandmark specification