Mining Optimization Laboratory

Transcription

1 Mining Optimization Laboratory Report One 2008/2009 Directed by Hooman Askari-Nasab School of Mining and Petroleum Engineering Department of Civil & Environmental Engineering, University of Alberta, Edmonton, Alberta, CANADA

2 All rights reserved, all material in this report is, unless otherwise stated, the property of the Mining Optimization Laboratory (MOL). Reproduction or retransmission of the materials, in whole or in part, in any manner, without the prior written consent of the copyright holder, is a violation of copyright law. This first report may be distributed to other potential sponsors for the advertisement purposes. The report may be circulated and disposed at your discretion; however, the following copyright notice must be adhered to. Copyright 2009, Mining Optimization Laboratory Mining Optimization Laboratory sponsors may utilize and disclose the report material and software within their organization with no prior permission of MOL. Contact information for requests for permission to reproduce or distribute materials available through this report is: Hooman Askari-Nasab, Ph.D., P.Eng. Director of Mining Optimization Laboratory Assistant Professor of Mining Engineering Markin/CNRL NREF Building Department of Civil & Environmental Engineering University of Alberta, Edmonton, AB, Canada T6G 2W2 Phone: (780) Fax: (780) hooman@ualberta.ca (i)

3 Executive Summary I am pleased to announce the release of the first Mining Optimization Laboratory (MOL) annual research report. MOL is an industrial affiliates program at the School of Mining and Petroleum Engineering at the University of Alberta. The program has been launched in January I would like to thank the support of the founding sponsors, Newmont USA Limited and Suncor Energy Inc. in sponsoring this initiative. It is anticipated that the value generated to sponsors by the outcome of research would create more industrial awareness and excitement to support the research group in long-run with the focus on two major themes (i) risk-based mine planning and design and (ii) simulation optimization of mining systems. The mission of MOL research group is to (i) train highly qualified personnel, and (ii) to extend the industry best practice guidelines and standards in the above-mentioned areas. I am sure this report will demonstrate the value generated to our sponsors through research and will give a flavor of the quality of research and deliverables to the potential sponsors. As promised, research and training has been conducted and documented in two main categories, mine planning and design and simulation optimization of mining systems. This report is a deliverable that would rationalize sponsorship of MOL research. The students at the MOL are just starting their program and they are going through a training phase by taking courses and working on mine planning projects. However, I am satisfied with the research results documented in this report and foresee a successful research group with focus on far-sighted fundamental research addressing industry applicative needs. This report presents a number of interesting contributions towards developing the major components of a risk-based mine planning framework. Let s review the contributions in Report One by considering some of the main contributors. Hooman has worked on large-scale long-term open pit production scheduling using mixed integer linear programming (MILP) techniques (papers 101, 102, and 104). The MILP models have the capability of improving the net present value of the production schedules significantly comparing to the traditional methods. An optimal schedule is generated with all technical constraints such as: mining-fleet capacity, shovel limits, processing options, multiple processes, multiple elements, and blending and head-grade constraints for ore and deleterious materials. One of the most important features of the MILP techniques is availability of a measure goodness of the solution compared against the theoretical optimum. We investigated the shortcomings of the MILP models used for open pit production scheduling, particularly the inability to solve large-scale real-size mining problems. To reduce the size of the open pit production scheduling problem we introduced the concept of mining-cuts into the MILP formulations. Blocks within the same level or mining bench are grouped into clusters based on their attributes, spatial location, rock type, and grade distribution. In this study, we have developed, implemented, and tested four MILP formulations for large-scale open pit production scheduling. The models were implemented in MATLAB and TOMLAB/CPLEX environment. The developed models proved to be able to handle deterministic large-scale mine production problems. A case study was conducted on an iron ore mine with more than 170 thousands blocks within the (ii)

4 final pit outline over twenty planning periods. The case studies have proved how effective MILP formulations are in generating high value practical schedules and maximizing the recovery through consistent feed grade to the plant. Another ongoing research direction followed by Hooman is mine planning using intelligent open pit production simulator, IOPS (papers 105 and 106). The intelligent agent theoretical framework for real size mine planning was developed as his PhD work based on reinforcement learning algorithms. The intelligent agent framework employs Q-learning algorithm to maximize the net present value of the mining operation and is a powerful tool in generating feasible optimal push-backs. IOPS, is developed and implemented in Java and MATLAB and tested through a mining case study against Whittle software. Behrang has carried out a very interesting research work on quantifying the uncertainty transferred from geological modeling into mine planning (paper 103). First, sequential Gaussian simulation is used to generate fifty realizations of an oil sands deposit. An optimum final pit limits design is carried out for each SGS realization while fixing all other technical and economic input parameters. Afterwards, the long-term schedule of each final pit shell is generated. Uncertainty in the final pit outline, net present value, production targets, and the head grade are assessed and presented and compared against the kriged and post simulation models. The results show that there is significant uncertainty in the longterm production schedules, even though a buffer stockpile was considered in the study. Kwame is an assistant professor of mining engineering at the Missouri University of Science and Technology, Rolla, USA. His current research interests include modelling and optimization, and life-cycle assessment of mining systems. He remains an integral part of our research group particularly in mathematical programming and simulation modeling. Kwame presented work on cable shovel kinematics and dynamic modeling combined with formation resistance modeling (paper 203) and a linear programming model of the asphalt mix design problem along with a numerical algorithm to solve the model (paper 205). Eugene, Samira, and Yashar have taken very important training steps toward their degrees with taking courses in simulation of industrial and engineering systems, and engineering optimization. Mixed integer programming formulations occasionally create a scattered block extraction order that cannot be implemented in practice. Yashar has made meaningful contributions in examining different techniques to improve the practical feasibility of scheduling patterns. Yashar has also documented the steps required to generate a short-term schedule on a mining bench using MIP models with Excel Solver (paper 110). Eugene and Samira joined us in January 2009; they have made meaningful contributions in pit optimization and blending considering the short time that they have been with MOL (papers 107 and 108). Another training/research activity undertaken by Eugene and Samira, was orebody modeling, optimization and pit design of an oil sands deposit in the McMurray formation (papers 111 and 112). They will continue working on geostatistical modeling of oil sands deposit as an integral part of their theses. Mohammad, Hesam, and Mahdi have joined us as PhD students in September 2009 with industrial engineering background. Thanks for joining us! September 2009 (iii)

5 First Annual Research Report of the Mining Optimization Laboratory (MOL) 2008/2009 Paper Page Title Table of Contents 100 Mine Planning and Design Mixed integer linear programming formulations for open pit production scheduling, Hooman Askari-Nasab and Kwame Awuah-Offei Numerical modelling of the MILP formulation for open pit production scheduling, Hooman Askari-Nasab and Kwame Awuah-Offei Transfer of geological uncertainty into mine planning. Behrang Koushavand, Hooman Askari-Nasab Modeling variable pit slopes in the open pit production scheduling MILP formulation, Hooman Askari-Nasab and Kwame Awuah-Offei An agent based framework for open pit mine planning, Hooman Askari- Nasab and Kwame Awuah-Offei Open pit optimisation using discounted economic block values, Hooman Askari-Nasab and Kwame Awuah-Offei Open pit limits optimization using linear programming, Eugene Ben-Awuah and Hooman Askari-Nasab Optimization of blending process in mine production, Samira Kalantari and Hooman Askari-Nasab Optimizing block extraction sequence with MIP method and investigating the effect of road condition on truck cycle time, Yashar Pourrahimian and Hooman Askari-Nasab A review of open pit mine production scheduling and a guide for using Excel Solver in modeling MIP problems, Yashar Pourrahimian and Hooman Askari-Nasab. (iv)

6 McMurray oil sands pit design using Gemcom and Whittle, Samira Kalantari and Hooman Askari-Nasab Orebody modeling, optimization and pit design using GEMS and Whittle, Eugene Ben-Awuah. 200 Optimization & Simulation Simulation of a crushing process, Eugene Ben-Awuah and Hooman Askari- Nasab Simulation of mine production using MATLAB and AweSim software, Samira Kalantari and Hooman Askari-Nasab Cable shovel dynamic simulation application in pre-blasting decisionmaking, Kwame Awuah-Offei and Hooman Askari-Nasab Reliability analysis of truck-shovel systems in mining, Kwame Awuah-Offei and Hooman Askari-Nasab Asphalt mix design optimization for efficient plant management, Kwame Awuah-Offei and Hooman Askari-Nasab Life cycle assessment of belt conveyor and truck haulage systems in an open pit mine, Kwame Awuah-Offei, David Checkel, and Hooman Askari- Nasab. 300 Software Related Guidelines for using TOMLAB on Linux clusters, Hooman Askari-Nasab. (v)

7 Mining Optimization Laboratory (MOL) Researchers / Graduate Students Following are researchers and students affiliated with Mining Optimization Laboratory in September Hooman Askari-Nasab Assistant Professor and Director of MOL 2. Kwame Awuah-Offei Assistant Professor of Mining Engineering 3. Mahdi Badiozamani PhD Student (new) /09 4. Eugene Ben-Awuah PhD Student /01 5. Hesameddin Eivazy PhD Student (new) /09 6. Yashar Pourraimian PhD Student /09 7. Samira Kalantari Msc Student /01 8. Behrang Koushavand PhD Student (CCG) /09 9. Mohammad Tabesh PhD Student (new) 2009/09 (vi)