Dynamic modelling of 3D stockpile for life-cycle management through sparse range point clouds Shi Zhao a , Tien-Fu Lu a, ⁎, Ben Koch b , Alan Hurdsman b a School of Mechanical Engineering, The University of Adelaide, SA 5005 Australia b MatrixGroup, Level 5 185 Victoria Square, Adelaide SA 5000, Australia abstract article info Article history: Received 22 October 2012 Received in revised form 5 August 2013 Accepted 28 September 2013 Available online 9 October 2013 Keywords: Stockpile management 3D modelling Laser Least square Bulk material Geometric stockpile models are always used to estimate the quantity and quality of iron ore in stockpile manage- ment systems. However, most of these models are not suitable for quality control purposes in iron-ore export fa- cilities because no measurement device is used to continuously provide updated information for the modelling. This paper presents a solution to accurately measure profiles along the stockpile surface in real time and automat- ically models the 3D stockpile during stacking and reclaiming operations. Using a 2D laser range finder, the spa- tial data of a stockpile, such as the positions of different layers, the cutting surfaces during the reclaiming are represented by mathematical equations. Experiments conducted in a laboratory environment indicated good re- sults and proved that the modelling approach is accurate and efficient. By making available the latest shapes of stockpiles and quality information, it is possible to predict the tonnage and quality grade with a greater degree of accuracy. Also, together with the reclaimer position data, it is possible to develop adaptive reclaiming algo- rithms to continually adjust the quality with respect to the quality requirement, which warrants meeting the re- quired quantity and quality while enhancing the export efficiency and increasing productivity. © 2013 Elsevier B.V. All rights reserved. 1. Introduction Statistical data indicate that 92% of iron ore mined in Australia is exported to global customers for steel making (Australian Bureau of Ag- ricultural and Resource Economics and Sciences, 2010). In such interna- tional trading, the quality of the shipped product is governed by two factors: the tenor of the iron-ore body and the short-term grading con- sistency. The latter one is the main criterion and is closely monitored by customers on a ship-by-ship basis. To control the product quality and meet customers' specifications, stockpiles are widely used in iron-ore handling due to their abilities to blend ores and homogenize quality grades of delivered products. The blending and homogenizing function is currently achieved by stacking iron ore into different geometric shapes and then reclaiming across these shapes slice-wise. Fig. 1 illustrates the ideal cross-sections resulted by three stacking methods. Out of the three, chevron stacking is very popular because of its simplicity and convenience. It is formed by moving the stacker forwards and backwards over the predefined length at an almost constant speed. The boom lifts a little at each end. Ide- ally, poured materials from the end of the boom form a thin layer with a triangular cross-section. Layers are stacked on top of each other after the zigzag motion is completed. When a chevron stockpile is recovered using a Bucket Wheel Reclaimer (BWR), which has a rotating wheel with buckets mounted on the tip of the boom, perpendicularly to the stacking direction, materials in these layers are simultaneously scooped up by buckets and completely mixed inside the rotating wheel. Consequently, the grade variability of reclaimed iron-ore is reduced. In order to obtain a uniform composition, a stockpile may have over one hundred thin layers. Or, the uniformity of a stockpile can be improved by repeating stacking and reclaiming operations several times. Because both stacking and reclaiming aim to minimize the variation in the quality grade, they are collectively referred to as blending in this paper. According to blending operations, a stockpile always has four phases across its life-cycle, as shown in Fig. 2. These phases are stacking, awaiting reclaiming, reclaiming and awaiting stacking. If all blending activities throughout the stockpile life-cycle are based on accurate qual- ity information of ore-bodies and real geometric shapes of layers, it will lead to an effective and efficient quality control plan. However, the qual- ity data with the highest degree of accuracy cannot be available in a real time framework due to the limitation of current sampling techniques. There always exists two to three hours in delay to obtain the assay result. More details on these sampling techniques and hardware are covered by Gerlach et al. (2002) and Petersen et al. (2004). Neverthe- less, delivered iron ore has to be processed and stacked to save storage space. That is to say, in the stacking phase, operations may have to be based on prior quality information from mining sites, which may be out of date. Similar problems also occur in the reclaiming phase. To maintain a relatively constant quality grade of delivered iron ore, reclaiming operations have to frequently suspend and wait for the International Journal of Mineral Processing 125 (2013) 61–77 ⁎ Corresponding author. Tel.: +61 8 8303 3556. E-mail addresses: shi.zhao@adelaide.edu.au (S. Zhao), tien-fu.lu@adelaide.edu.au (T.-F. Lu), benk@matrixgroup.com.au (B. Koch), alanh@matrixgroup.com.au (A. Hurdsman). 0301-7516/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.minpro.2013.09.009 Contents lists available at ScienceDirect International Journal of Mineral Processing journal homepage: www.elsevier.com/locate/ijminpro