Slurry Transport System of the ROBOMINERS Prototype
1 ROBOMINERS – project overview
In the realm of underground mining, the challenging conditions and potential dangers arising from events like rock fall or bursts consistently pose inherent risks to on-site personnel. Consequently, there is an ongoing shift towards complete mechanization and subsequent automation of mining processes. Nevertheless, certain tasks prove resistant to automation, necessitating human presence in potentially perilous areas for underground maintenance. Presently, there is preliminary research and development focused on the development of robots anticipated to replace human labor in underground mining within the next three decades. Future challenges in the mining industry, influenced by considerations of sustainability and ecological impact, demand additional dedication to research and development. Through the integration of fully automated machines or autonomous robots, the possibility of opening new deposits or economically operating previously closed mines becomes feasible. Potential roles for robots in mining encompass machine maintenance, exploration of abandoned mines and mining activities, particularly in difficult-to-access areas. The design of autonomous robots may vary significantly based on factors such as mine layout, mining method and deposit type, deviating from the structure of current machines. (1, 2)
The concept driving the ROBOMINERS project is the exploration of innovative technologies for accessing challenging, untapped deposits. In this four-year initiative, 14 institutions across eleven EU countries collaborated to establish critical knowledge about future technologies employed in robot-operated mining scenarios. The goal was to develop and test a first-generation prototype of a small-scale mining robot. It is crucial to note that while this project and the resultant small-scale mining robot are intended for feasibility studies exploring various technologies for scenarios requiring excavation of small volumes, they are not intended to replace conventional mining and tunnelling equipment. Instead, the outcomes of the project aim to provide an overview of technologies that can be incorporated into future mining robots. (3, 4)
2 ROBOMINERS RM1 prototype
The largest of the several ROBOMINERS prototypes (RM1) is a full-scale robot, equipped with perception tools, an excavation tool and a material transport system. The key functionalities have been tested in summer 2023 in an Estonian open-pit oilshale mine. Further testing of the perception equipment has been performed in an underground mine in Slovenia. The prototype is fully water-hydraulically powered with a total power of 30 kW and an approximate weight of 1,350 kg, tethered, remote-controlled and with an overall length of approximately 4,500 mm. In the next section, the core modules of the RM1 prototype are presented.
2.1 Design
The body module comprises a structural hull and associated subsystems, including actuators responsible for the movement of the screw units. The diameter of the robot, exclusive of screws, measures 800 mm (expanding to 1,000 mm with screws extended). (5, 6)
For enhanced traction capabilities, the prototype will incorporate two main modules. This additional weight is essential to counterbalance the forces exerted by the production tool during excavation operations. Moreover, this modification facilitates a more comprehensive examination of the modular structure, its functionality and capabilities within an authentic operational setting. Both the first and second modules share identical dimensions and feature the same main component design, although their structural designs differ. The primary distinction lies in the fact that the second module is equipped with only two screw units. This alteration introduces additional free volume, creating more space for system components. Figure 1 illustrates the final design of the RM1 prototype. (7)
2.2 Locomotion system
The locomotion system of the ROBOMINERS comprises a drive screw and a leg, which functions as a movable auxiliary structure situated between the screw and the main hull. The screw mechanism is rooted in the Archimedes screw principle, where the installed units are powered by water-hydraulic motors, and the leg’s movement is driven by a water-hydraulic actuator. The incorporation of two independently driven screws allows for exceptionally small turning radii.
The actuation design principle remains consistent across all modules within the ROBOMINERS prototype. A water-hydraulic system with an open circuit is employed to power the actuators. Opting for water as the pressure medium presents notable advantages over oil, including high availability, low cost, environmental friendliness and non-flammability. However, as water serves as the hydraulic medium, considerations must be made for potential challenges such as corrosion, low lubrication properties and an elevated risk of leakage between moving parts in the design of the actuation system. The screw units, four of which are attached to Module 1 and two to Module 2, include the screw (1) housing the water-hydraulic screw drive unit, leg radial actuators (2) and leg side actuators (3). Figure 2 provides a visualization of the structure of the screw unit. (7)
2.3 Production tool
The part-face cutting technology was selected for various reasons, including its capability for continuous rock mining, adaptability to changing rock conditions, the robust nature of the mining tool and its universal applicability in both dry and submerged conditions. However, it is essential to acknowledge the drawbacks associated with this technology. The efficiency of mechanical cutting systems is significantly influenced by the strength of the rock being excavated. High reaction forces are typically generated during the cutting process, necessitating absorption by the machine or robot. (8, 9)
Integral to the production tool’s development is the evaluation of its excavation capabilities. Consequently, a production tool test rig (Figure 3) has been meticulously planned and constructed in the laboratory to systematically assess the performance of the production tool. To ensure the effectiveness of the test rig, the boundary conditions and power requirements of the part-face cutter head were initially defined. (9)
Various concrete specimens were created, including those with compressive strengths of 20 and 30 MPa UCS (Uniaxial Compressive Strength). Additionally, concrete specimens featuring oilshale (16 MPa UCS) and limestone (approximately 60 to 80 MPa UCS, not measured) inserts were fabricated. (9)
Key performance parameters were defined, including the excavation rate. In very soft rock, such as oil shale, the excavation rate reached approximately 1 m³/h. In stronger and more compact rocks, such as concrete with a UCS of 30 MPa, the excavation rate decreased to 0.2 m³/h, and the maximum UCS for this production tool was deemed to be 40 MPa. Beyond this limit, cutting operations are considered uneconomical. The scalability of part-face cutter heads is not a significant obstacle, but their capabilities diminish significantly at smaller scales. It has been demonstrated that, up to a certain rock strength, this method is applicable but not economically viable. In hard rock conditions, future challenges demand different approaches, such as alternative or combined excavation methods or the traditional drill and blast technique. (9)
2.4 Solid-state and mid-range perception sensors
In the context of the ROBOMINERS project, a novel approach was developed for autonomous measurements conducted by a robotic platform. In this test scenario, a robotic arm serves as an analogue to the prototypes’ boom, where a sensing device can be mounted instead of the production tool. At the head of the robotic arm, a platform accommodating either a line of eight electrodes or a cluster of 32 electrodes (arranged in 8 x 4 lines) is used for linear or 2.5D Electrical Resistivity Tomography (ERT) and Induced Polarization (IP) measurements. (10)
The experiments involved two types of electrodes. The first type featured a screw with a 3D-printed cone filled with a mixture of graphite and silicon. The second type employed spring-mounted pogo-pins with a layer of copper and silver at their ends (Figure 4). This latter electrode design was developed to adapt to and compensate for the subtle changes in the topography of the mine wall being mapped. (10)
2.5 Slurry transport system
2.5.1 Design criteria
The design criteria for the ROBOMINERS prototype ore transport system were derived from the RM1 prototype design status in early 2022. At that time, the layout of the RM1 prototype, including locomotion system and boom design, had all been established. Figure 5 presents the RM1 prototype design which formed the basis for the ore transport system design criteria.
Source/Quelle: MU Leoben
The main governing criterium followed the limitations on size and form factor. As can be seen from figure 5, a very limited space for an ore collection system near the production tool was available without impeding manoeuvrability of the boom and/or locomotion of RM1.
A second governing criterium was the capacity and the expected grain size of the excavated ore. With the design of the production tool established, the required capacity of the RM1 ore collection and transport system was limited to the cutting capacity of the tool. Laboratory testing of the tool established that up to 1 m³/h of intact ore rock per hour operation could be extracted. This testing further established a grain size distribution for the excavated material with a D50 of approximately 2 mm and a D90 of 10 mm. These values were taken as a baseline for the RM1 ore transport system during development. Transport distance was determined by the expected RM1 demonstration layout and was set at 10 to 20 m between the prototype and the Laser-induced Breakdown-spectroscopy (LIBS) system.
Other design criteria were low technical complexity, to not further exacerbate the complexity of the prototype, and ease of installation and repair during the demonstration of RM1. As the time left in the ROBOMINERS project was limited, a form of rapid prototyping or application of off-the-shelf parts was necessary to achieve the project goals on time and within budget.
2.5.2 Concept development
Early technical concepts for an ore transport system suitable for the RM1 prototype were based on both mechanical and hydraulic ore collection and transport systems. Considering the design of RM1, an ore collection system was best positioned underneath the production tool for two reasons: space availability and chance of picking up material excavated by the production tool.
Further exploring industrial analogues; the Venturi-based pump was identified as a potential solution. A Venturi-based pump does not use moving parts and is driven by water. Venturi pumps can be designed and manufactured in a small and relatively flexible form factor. Examples of Venturi-based pumps are extractor pumps and powder mixers as shown in figure 6.
2.5.3 Venturi system development
Having established the basis of design for the RM1 ore collection and transport system, the development and manufacturing stage was initiated. With limited time available it was decided to use 3D printing technology for rapid prototyping of the Venturi pump, auxiliary parts and mounting brackets. Initially, the plan called for prototyping by 3D printing in plastics after which a final design could have been 3D printed in metal or produced by other manufacturing methods. During the rapid prototyping phase however a vast amount of experience was gained by the design team with 3D printing. The potential strength of the plastic parts that can produced by 3D printing when carefully selecting printing parameters and controlling the printer environment was deemed sufficient to withstand the expected physical forces on the Venturi system during the planned demonstration. This assumption was proven correct during two weeks of testing in real-life conditions of the RM1 prototype in an opencast mine in Estonia in the summer of 2023.
The basic design of the developed Venturi pump for the RM1 prototype was based on the work by Xu et al. (11), which is shown in figure 7.
The annular jet pump (AJP) is in large parts defined by three parameters, namely the suction chamber angle α, diffuser angle β and a factor m for the ratio between the cross-sectional areas of the secondary and primary flows (11). Xu et al. (11) investigated the optimal value for m, i. e. achieving the highest efficiency, arriving at a value of 2,11 with α ≈ 25° and β ≈ 5°. Based on these parameters several designs of the RM1 AJP were prepared, as shown in figure 8, and subsequently tested for attainable vacuum at the suction inlet.
The performance of the early designs was tested using a multi-stage centrifugal pump with a 6 m³/h capacity at ~ 6 bar maximum pressure. It was found that the design parameter for m with values around 2 to 3 was not resulting in sufficient suction, i. e. vacuum at the secondary inlet at the given motive medium conditions (5 to 6 m³/h at 5 to 6 bar).
The version 7 with the independent nozzles was found to perform better, i. e. achieving a higher vacuum at the secondary inlet, than the version 8 at the stated test conditions of 5 to 6 m³/h at 5 to 6 bar primary inlet pressure. As the date of demonstration of RM1 was approaching rapidly, it was decided that version 7 was to be used during the demonstration of RM1 and version 8 was to be used for further testing (after the demonstration) with an inlet flow rate of up to 10 m³/h at a pressure of 10 bar. To allow for the higher pressure to be applied to version 8, a special high-pressure version of the AJP design was prepared by increasing overall wall thickness. All the versions just described are presented in figure 9.
2.5.4 Venturi system – manufacturing and integration
Once the principal design of the RM1 AJP was established, an integration of the AJP into the RM1 boom layout was initiated. The goal of the integration was to allow mounting the AJP on the underside of the boom while attached to a suction nozzle with an integrated screen for classification based on grain size. The classification was established using perforated plates with a range in hole diameter from 10 to 16 mm. As the space underneath the boom is limited, a compact and rounded form factor was sought for the AJP and suction nozzle. Lastly, a mounting system for the AJP and nozzle assembly was also designed and prepared. The developed mounting system is modular allowing the interchanging of parts to enable mounting sensors or auxiliary water nozzle to support excavation.
The use of two side-by-side Venturis enabled a large suction nozzle surface covering the full width of the boom. The two Venturis have a common water inlet and each has its own diffusers from which the output is combined into a single large diameter hose leading to the back of the prototype. The mounting system allows for some flexibility in the (axial) position of the AJP along the boom thus allowing variation of the distance between the suction nozzle and the production tool. The overall design process of the integrated AJP, nozzle and mounting system took a few iterations. The final version is presented in figure 10.
2.5.5 Testing and demonstration
Testing and demonstration of the RM1 prototype took place in the Kunda oil shale mine near the city of Rackvere in Estonia. The ore collection system comprising the suction nozzle, with a 16 mm holed plate, the double Venturi AJP and the mounting system were successfully operated during the demonstration. The fully assembled ore collection system as used during the demonstration is shown in figure 11. The top of the mounting system has been equipped by a separate sensor system comprising two compressed air-driven pistons.
The ore cut by the production tool during the demonstration is transported to the back of the robot by the Venturi-based slurry pump. This pump is driven by medium pressure water (+/- 5 to 6 bar) flowing through a Venturi and consequently creating a vacuum. A multistage centrifugal pump taking in water from the main water basin was used to provide the medium pressure water flow needed to drive the Venturis. The vacuum created is exerted on a suction nozzle with a holed plate on the underside of the boom, directly behind the production tool.
The system was able to successfully suck up rock particles up to 5 mm and transport them to the back of the robot via the LIBS system (total distance exceeding 20 m). The Venturi pump was unsuccessful in extracting all the cut rock by the production tool as the grain size after cutting was too large and could consequently not pass the sieve on the Venturi nozzle. This was expected considering the faulted nature of the ore rock at the Rakvere site. The cutterhead has been able to cut rock to the desired grain size in the laboratory, albeit this was on unfractured samples thus enabling a consistent small particle size for the excavated material.
2.5.6 Future development
To enable the Venturi to extract all the ore rock cut by the production head, the grainsize must be consistently smaller and/or the passing size of the Venturi system needs to be larger. Another improvement would be to use a hollow production tool which allows material to be sucked through the cutter head and into a concentric Venturi system just behind the cutter head. In this way, cut ore rock will mostly be sucked directly into the ore collection system rather than falling on the floor. Secondary Venturi pumps/suction nozzles could be installed near the floor in front of the robot’s locomotion system to ensure an ore-free path for the robot to move forward on. A rotary crusher can be integrated into this design to ensure a suitable grain size distribution and consequently full extraction of all ore rock excavated by the production tool.
Notwithstanding the shortcomings of the ore collection system developed for the ROBOMINERS prototype RM1, Venturi pumps are believed to be relevant for robotic mining in the future due to their simplicity. This simplicity is created by the lack of moving parts in the Venturi pump and the ability to use water as both the pump’s driving force and the transport fluid for a slurry. In submerged robotic mining situations, a Venturi-based pump could drive a very reliable ore collection system without the need for extensive maintenance and the possibility to make (simple) components onsite using 3D printing technology rather than having to order complex parts from a warehouse (far) offsite.
2.6 Mineralogical segment sensors
The LIBS and X-Ray Fluorescence (XRF) techniques exhibit advanced elemental measurement capabilities and are widely used, including as cross-belt analyzers in mineral processing. LIBS, though limited in homogeneity assessment, offers a straightforward spectroscopic setup and broader elemental detection capabilities compared to XRF, especially for elements with atomic numbers below 20. The growing interest in LIBS for geological analysis arises from its simplicity, cost-effectiveness, simultaneous multi-elemental analysis and extensive elemental detection. The wealth of information present in each LIBS spectrum, with thousands of emission peaks, poses challenges in real-time and reliable data processing. However, advancements in computing power and dedicated machine learning methods are improving the efficiency of real-time LIBS data treatment, making the interpretation of mineral contents from elemental LIBS results more reliable (as exemplified in (12)). (13)
In July 2023, a field demonstration took place at the Kunda mine, showcasing the operation of the slurry analyzer integrated with the ROBOMINERS RM1 prototype. The Kunda mine exploits oil shale deposits in an open-pit configuration. The purpose of this demonstration was to functionally illustrate how the LIBS slurry analyzer (Figure 12) could provide a qualitative output of the extracted material by RM1. The field tests were successful, with continuous acquisition of LIBS spectra during RM1 operation. A special software was adapted for the demonstration, displaying the raw peak heights of a few elements present in the extracted material, including silicon (Si), calzium (Ca), magnesium (Mg), aluminum (Al) and iron (Fe). The detected iron likely originated from pyrite (FeS2) occurrences in the oil shale layers. (13)
3 RM1 field test
The purpose of the field tests and demonstrations was to both test and publicly demonstrate capabilities of the ROBOMINERS RM1 prototype and sensors developed in the project. Following the successful full commissioning of RM1, it was relocated to the designated area for further testing and demonstration (Figure 13).
The subsequent testing phase focused on validating the performance of the production tool module and locomotion. Control algorithms for both the production tool module and locomotion functions were refined during testing to enhance their responsiveness under real-life non-ideal conditions. RM1 was tested successfully and presented in public demonstrations to different audiences. Demonstrations were performed in dry conditions as well as with the robot partially submerged. (14)
Funding
This project has received funding from the European Union’s Horizon2020 research and innovation programme under grant agreement No. 820971.
References / Quellenverzeichnis
(1) Hiltz, R. (2020): Taking a step into the robotic future. URL: https://www.miningmagazine.com/technology/news/1387411/taking-step-robotic-future (visited on 19/12/2023).
(2) Siciliano, B.; Khatib, O. (Hg.) (2008): Springer Handbook of Robotics. Springer Berlin, Heidelberg. Springer Handbooks. ISBN 9783540303015.
(3) Berner, M.; Sifferlinger, N. A. (2021): H2020 – ROBOMINERS. In: BHM Berg- und Hüttenmännische Monatshefte 166.2, S. 59 – 63. ISSN: 0005-8912. doi: 10.1007/s00501-020-01074-y.
(4) European Commission (2018): ROBOMINERS Proposal SEP-210520664. Resilient Bio-inspired Modular Robotic Miners / ROBOMINERS.
(5) Aaltonen, J. et al. (2020): ROBOMINERS deliverable D3.2: Robot conceptual design report. Ed. by ROBOMINERS Consortium.
(6) Aaltonen, J. et al. (2020): ROBOMINERS deliverable D3.1. Robominer requirement specification. Ed. by ROBOMINERS Consortium.
(7) Aaltonen, J. et al. (2022): ROBOMINERS deliverable D3.3. Robominer subsystems and components laboratory test report. Ed. by ROBOMINERS Consortium.
(8) Berner, M.; Sifferlinger, N. A. (2020): Analysis of Excavation Methods for a Small-scale Mining Robot. Proceedings of the International Symposium on Automation and Robotics (ISARC). Kitakyshu, Japan (Online Conference), pp 481 – 490. doi: 10.22260/ISARC2020/0067.
(9) Berner, M. (2023): An Applicability Assessment of Excavation Tools for Small-scale Mining Robots. PhD thesis. Montanuniversität Leoben, 2023. doi: https://doi.org/10.34901/mul.pub.2023.159.
(10) Stasi, G. et al. (2023): ROBOMINERS deliverable D6.2. Solid-state and mid-range perception sensors. Ed. by ROBOMINERS Consortium.
(11) Xu, K.; Wang, G.; Wang, L. et al. (2020): Parameter Analysis and Optimization of Annular Jet Pump Based on Kriging Model. In: Appl. Sci. 2020, 10, 7860. doi: 10.3390/app10217860.
(12) Baele, J.-M. et al. (2021): Trace-element imaging at macroscopic scale in a Belgian sphalerite-galena ore using Laser-Induced Breakdown Spectroscopy (LIBS). In: Geologica Belgica, 24/3–4: 125 – 136. doi: 10.20341/gb.2021.003.
(13) Burlet, C. (2023): ROBOMINERS deliverable D6.3. Mineralogical segment sensors. Ed. by ROBOMINERS Consortium.
(14) Aaltonen, J. et al. (2023): ROBOMINERS deliverable D7.3. Field test report. Ed. by ROBOMINERS Consortium.
