Abstract
Iron ore tailings (IOT) from tailings dams cause significant environmental, economic, and social impacts, which has motivated the search for reuse alternatives. In this scenario, the present work uses reverse logistics to size the resources required to reuse the IOT for road infrastructure. A reverse logistics system was modelled to transport the IOT from the tailings dam to a processing plant and then to the construction site of a road. Resource capacity constraints and operational limitations were considered, and the economic feasibility of the system was analysed. Two scenarios for tailings transportation on the dam/plant route were simulated aiming at a cost less than R$ 25 (US$ 6.13) per tonne: (1) trucks; and (2) conveyor belts. The economic feasibility of both scenarios was proven, with scenario 1 presenting the most promising results – a transport distance of 290 km within the established price limit. This methodology can thus be applied to encourage the large-scale reuse of IOT.
Introduction
In 2017 alone, Brazil extracted 585 million tonnes of iron ore, being the world’s third-largest exporter (Ministry of Mines and Energy, 2018). According to the Ministry of Mines and Energy (2018), this ore had an average iron content of 54%. The increase in iron ore extraction in recent years makes the processing of poorer ores (with lower iron content) inevitable (Brazilian Institute of Mining, 2016). This fact, in turn, has led to increased tailings generation, making their final disposal and management practices increasingly important.
The large volume of tailings results in a spatial problem (due to the vast area required for storage), and a temporal problem, since it takes long periods for the tailings to be managed and rehabilitated (Adiansyah et al., 2015). Conventional iron ore tailings (IOT) typically range between 30 and 50% solids (Franks et al., 2011). Due to their fluid rheology, tailings dams are the typical destination of this waste in some countries. These dams are usually built using the coarse fraction of the tailings themselves, with steep slopes and successive heightening (Ávila, 2012; Gomes et al., 2016).
With the advancement of mining activities and the increased scale of operation, many of these dams began to present structural risks (Ávila, 2012). As a result of poor management and monitoring practices, two major dam collapses occurred in Brazil in 2015 and 2019. These disasters led to the release of millions of tonnes of IOT, several victims, and significant impacts on properties and ecosystems (Burritt and Christ, 2018). Similar accidents caused major damages in other countries as well, especially the People’s Republic of China (Wei et al., 2013).
In this scenario, various researchers have focused on the reuse of the IOT as aggregate for construction (Almeida et al., 2020; Fontes et al., 2016; Mendes et al., 2019b; Sant’ana Filho et al., 2016), pigment (Fontes et al., 2018; Galvão, et al., 2018), raw material for ceramics (Behera et al., 2019; Fontes et al., 2019; Mendes et al., 2019a), and as road infrastructure (Bastos et al., 2016; Ojuri et al., 2017). These studies show that there is great technical potential for the rehabilitation of IOT in the civil construction and road production sectors. Hence, to ensure that these solutions evolve from laboratory scale, it is imperative to develop logistics systems to support them. Since IOT is a residue from the mining industry, a reverse logistics system would be the most appropriate strategy.
Reverse logistics systems are practical solutions adopted by companies to implement reverse flows, consequently achieving economic and environmental gains (Couto et al., 2017; Kinobe et al., 2012). The reverse logistics field nowadays incorporates waste prevention, reuse, recycling, recovery, and correct disposal (Kinobe et al., 2012). It considers parameters such as generation rates, composition, collection and storage, enabling the cost-effective treatment of residues through proper environmental practices (Oyola-Cervantes and Amaya-Mier, 2019).
While the broad field of waste management focuses on the collection and treatment of waste in general, reverse logistics deals with residues that have some value to be recovered (Kinobe, et al., 2012; Veiga, 2013). It aims at improving the resale value of the returned products, using them as parts or extracting materials to employ in new products (Agrawal and Singh, 2019). To this purpose, these products can be directed back into the same production chain or to the production chains of other industries/companies.
A comprehensive and integrated approach to the design and implementation of reverse logistics systems has yet to be established. Among the available literature, Dowlatshahi (2005) proposes a strategic decision-making process following four stages: (a) meet customers’ requirements; (b) understand current regulatory and environmental issues, as well as future trends and possible changes; (c) implement a profitable operation; and (d) ensure that the quality of the new product meets or exceeds that of similar virgin products. Reverse logistics systems are applied in many fields, such as electronic waste (Ahluwalia and Nema, 2006; Guarnieri et al., 2016), construction and demolition waste (Oliveira Neto and Correia, 2019), scrap tires (de Souza and D’Agosto, 2013; Oyola-Cervantes and Amaya-Mier, 2019), and post-consumer packaging (Couto et al., 2017; Veiga, 2013).
In this scenario, the present work proposes a reverse logistics system to transport the IOT stored in a tailings dam to the construction site of a road, including the system’s configuration, equipment, and set up. It also compares the advantages and limitations of two IOT transport methods: conveyor belts; and dump trucks. The methodology addresses both the resources sizing and the economic analysis of an emblematic example, inspired by the operation of typical iron ore mines in Brazil.
It is noteworthy that, to the best of our knowledge, no literature was found applying reverse logistics principles to the reuse of IOT after their disposal in tailings dams. Given that this disposal causes major economic and environmental issues, the development of efficient and cost-effective processes for their mass-scale reuse is imperative. Additionally, the proposed resource sizing method can be applied to the reverse logistics systems of other minerals and industries. In this sense, this process can be used to support both tactical and strategic decisions for various reuse projects.
Materials and methods
Description of the proposed reverse logistics system
The present work addresses a hypothetical demand from a fictitious mining company to allocate the IOT from one of its dams for reuse in road paving. The operations of a prominent Brazilian multinational mining company were taken as a basis for this study. This company runs extraction and logistics activities and is one of the largest mining companies in the world.
The present work adopted it as a benchmark for two main reasons. Firstly, the company is obligated to meet the recommendations of Public Civil Inquiry 1.22.020.000246/2015-34 (Federal Public Ministry, 2016), published after the dam collapse of Fundão in 2015, which requires that all mining companies reuse at least 70% of their IOT before 2025. Secondly, the approval of research proposal TEC-RDP-00201/2010 (Minas Gerais State Agency for Research and Development – FAPEMIG/VALE S.A., 2010) allowed the research team to access the company’s facilities. This possibility, in turn, enabled data collection and diagnosis, and the proposal of solutions to contribute to the better management of IOT, such as the present reverse logistics system.
To pursue this objective, a discrete simulation model was developed (Banks et al., 2014), allowing inferences concerning the operating characteristics of the proposed reverse logistics system. It comprises: (a) the removal of the tailings stored in the dam; (b) the transportation of this material to the processing plant, where it will be processed and stored in stockpiles for 48 hours; and (c) the transportation of the tailings to the site where the road will be built. Figure 1 outlines the proposed reverse logistics system.
Proposed reverse logistics system to transport the iron ore tailings (IOT) from the tailings dam to the road construction site.
The economic feasibility for each configuration of the proposed logistics system was analysed through a preliminary economic analysis. The longer the distance from the stockpiles to the construction site, the higher the number of trucks required to maintain a constant flow of IOT to build the road. Each truck has costs associated with acquisition, operation, and maintenance. On the other hand, the higher the extension achievable within the cost restrictions, the more IOT will be reused and the greater will be the impact of the reverse logistics system to the region.
In this sense, a sensitivity analysis was performed to investigate the optimum distance between the processing plant and construction site. The distance was chosen as a limiting factor since it strongly influences the economic feasibility of the proposed logistics system, also determining the effective area in which the IOT can be reused.
The optimum distance was determined from the Valuation Method of Transport on Demand, which considers the maximum transport cost that matches the total cost of the disposal of tailings (Araújo et al., 2014). The parameters – type of vehicle, delivery time in a continuous flow, operating costs, seasonality, roads characteristics, and depreciation of the equipment – were kept fixed so that the distance was the main factor influencing the transport costs.
To evaluate the proposed system, a simulation was performed. The objectives of the simulation are:
(a) To calculate the number of resources (excavators, loaders, and trucks) needed to meet the system demand.
(b) To identify the operational constraints in the IOT transport system, from the removal from the dam to the arrival at the road construction site.
Two scenarios were evaluated regarding transport alternatives on the dam/plant route:
Scenario 1: dump trucks transporting the IOT.
Scenario 2: conveyor belts carrying the IOT.
The choice of these scenarios was based on the regional availability of transport modes and the cost of their implementation. In other words, the equipment and operations in the proposed reverse logistics system are the same as in the forward flow of mining companies (see, for example, Afrapoli et al. (2019); Rodrigues and Pinto (2012)). The similarity was purposefully established to reduce the constraints when implementing the proposed reverse logistics system in real scale.
The activities considered in the logistics system take place in three different locations: the tailings dam; the processing plant; and the construction site where the road is being built. The following subsections will show the inherent specificities of each of these locations, a description of the activities performed, as well as the resources needed to carry out each activity.
At the tailings dam
The activities begin at the tailings dam. IOT in the sludge state are removed by excavators, which immediately load the trucks (scenario 1) or conveyor belts (scenario 2). In the current system, a distance of 7 km between the dam and the processing plant was determined. This distance is necessary to ensure that the processing plant is not within the influence zone of the dam’s operations.
In scenario 1, as soon as the IOT are unloaded into the processing plant, the trucks are available to return to the dam and be reloaded by the excavators, initiating the first cycle of trucks, named dam/plant.
At the processing plant
The IOT arrives at the industrial processing plant, where it is processed and subsequently deposited in stockpiles with the assistance of a loader and three conveyor belts (lengths of 20, 60 and 100 m). The processing of the IOT consists of mixing it with an agent to improve rigidity and keeping it stored for 48 hours to stabilise the material. The costs of this process were included in the simulation. Subsequently, the loader carries the processed IOT into one of the conveyor belts, responsible for forming stockpiles with the tailings.
Each of these belts forms eight piles, containing 398 tonnes of tailings, each. Each pile takes 1 hour to be completed. Thus, the first belt works for 8 hours, carrying a total of 3184 tonne of IOT to the stockpiles. As soon as this amount is transported, the first belt is turned off and immediately the second belt goes into operation, starting the same process. After the second belt fills up eight piles, it is turned off, and the third belt begins filling up eight more piles. As soon as the third belt finishes forming the eight IOT piles, the first belt is turned on, and the process is resumed.
From the stockpiles to the road construction site
The IOT stored in the piles is removed by a loader that supplies the dump trucks, which are responsible for the transport to the road construction site. As soon as the first truck unloads the IOT on the set spot, it returns to the stockpiles and becomes available again to be loaded, initiating the second truck cycle, named piles/pavement.
Resources sizing
Conceptual model
As a result of these activities, the conceptual model was created. It translates the flow of real system activities into a sequential, analytical, and logical language, as shown in Figure 2.
Conceptual model of the reverse logistics system proposed.
Data collection
This step consisted of identifying the data that will supply the simulator. The input data are those coming from the specific characteristics of the studied system, shown in Table 1.
Input data.
Table 2 presents the probabilistic data adopted as the input of the simulation model. The loading/unloading times and speed of the equipment were obtained from their manufacturers’ catalogues (CAT, 2019; Volkswagen, 2019). Note that all parameters described in Table 2 use the normal probability distribution, where the averages were estimated based on the distances between the activities and on the time spent by the equipment to perform an activity. The value of 10% was assumed for the deviations, since the variability in this type of system is not high, as observed by Ribeiro et al. (2016).
Probabilistic expressions used as input parameters of the simulation model.
Model translation and validation
The developed conceptual model was translated into computational language so that the simulations could be performed. To this purpose, the Arena 11.0 software was adopted, using the Basic Process, Advanced Process, and Advanced Transfer templates. Each simulation used 7 days of warm-up period, ten replications, and a simulation run time of 180 days.
The simulation model was calibrated using an interactive method, in which the resources (i.e. the number of trucks, excavators, and loaders) were adjusted until the control data and system constraints were satisfied. System constraints included the distance between the dam site and processing plant, IOT storage time in the processing plant, capacity of the equipment, speed of the conveyor belts, and distance between the processing plant and the construction site, as shown in Table 1.
The control data (see Table 3) was evaluated at each simulation to provide proper resource sizing and ensure production volume. The amount of IOT in the system is given by the capacity of the entire set of equipment in each step of the process and was used to verify the equilibrium of the system.
Control data.
The resources sizing was based on equipment utilisation rates. The allowed utilisation rate of excavator, loader and truck resources was limited to 90% to consider random variations inherent in the system. As an illustration, we brought one of the iterations performed in scenario 1, considering a 50-km distance from the stockpiles to the construction site. This simulation stabilised assuming 14 22-tonne trucks in the dam/plant cycle (occupied 87% of the time), 21 40-tonne trucks in the piles/pavement cycle (occupied 85% of the time), 1 9-tonne excavator (occupied 75% of the time), 1 9-tonne loader (occupied 48% of the time), and 1 9-tonne stockpile loader (occupied 73% of the time). With this setup, the system processed and transported a total of 14,720,460 tonnes of IOT in a month.
No priority was established among the equipment – the amount of IOT in the system was limited by the first item that reached its maximum allowed utilisation rate (90%). The resource that mostly constrained the amount of IOT in the system was the trucks along the piles/pavement route. In this sense, a preliminary sensitivity analysis was carried out to determine how many trucks would be necessary according to this distance.
Economic feasibility
After the simulations were performed, an economic feasibility analysis was carried out. This analysis involves the costs of the resources calculated for each system location during the adopted lifespan of the logistics system.
Costs related to the acquisition of the trucks (22 tonnes and 40 tonnes) were provided by the Economic Research Institute Foundation, whch monthly publishes the average prices of vehicles in the Brazilian market (Economic Research Institute Foundation, 2019). The costs related to the acquisition of the excavator and loader, as well as the costs inherent to the operator of each equipment, were obtained from the National System of Costs Survey and Indexes of Construction (SINAPI). These tables summarise the costs and productivity rates of civil construction and geotechnical services in Brazil each month (Caixa Econômica Federal, 2019). Finally, the costs related to the acquisition and operation of the conveyor belts were obtained and adapted from the study of Ribeiro et al. (2016).
The IOT transported to the road construction site will replace the conventional aggregate used for infrastructure construction. According to the SINAPI tables for May 2018 (Caixa Econômica Federal, 2019), the average price of a tonne of sand is R$25.00 (US$ 6.13 1 ). Thus, the maximum cost of R$ 25.00 per tonne of transported IOT was adopted as a parameter to certify the economic viability of building the proposed logistics system.
The expected lifespan of the system is 20 years. This value was considered when calculating the acquisition, maintenance, labour, interest, fuel, and electricity costs. The total costs for each of the locations of the logistics system were obtained according to Table 4.
Logistics costs for both simulation scenarios.
The costs of loaders (CL1 and CL2), excavator (CEX1) and trucks (CDTD and CDTS) were calculated according to Equation (1), while the cost of conveyor belts (CCBD and CCBP) was given by Equation (2):
Finally, the distance between the stockpiles and the road construction site was varied in 20-km increments from an initial presumed distance of 30 km. For each variation, the corresponding cost was calculated. This procedure was performed in both scenarios to find the maximum distance equivalent to the cost of R$ 25per tonne of transported IOT.
Results and discussion
Resource sizing
Through the simulations, the resources of the model were suitably sized in the two proposed configurations (scenarios 1 and 2). As a result, the production per month achieved about 14.7 million tonnes of processed IOT in both scenarios. The final quantities of each resource considered in the reverse logistics system in each simulated scenario are presented in Table 5. The results showed that 14 trucks are required to achieve the performance of scenario 1 conveyor belt in the dam/piles cycle.
Quantity of each resource in each scenario.
Table 6 shows the number of trucks required to transport the stockpiled IOT to the pavement construction site for the various distances tested. It is observed that the number of trucks on the piles/pavement cycle is the same for both scenarios. This result is not impacted by the type of transport system when moving the IOT from the dam to the processing plant; it is only impacted by the amount of stockpiled IOT and the distance to where the road is being built.
Number of trucks required according to the distance between the stockpiles and the road construction site (route piles/pavement).
Economic feasibility
Finally, based on the costs calculated for each resource, Table 7 shows the cost per tonne of processed IOT reused through the proposed reverse logistics system. The simulations and economic analysis indicated that, in scenario 1 (trucking at the dam/plant cycle), a road being built up to 290 km from the processing plant has economic advantages over purchasing natural aggregates. In contrast, in scenario 2 (conveyor belts in the dam/plant cycle), the limiting price of R$ 25 is obtained only up to a maximum radius of 170 km.
Transport cost per tonne of processed iron ore tailings.
The bold values mark the greatest distances in which the cost of transport was below R$ 25.
In short, the reverse logistics systems scaled in both scenarios are economically viable, since they found considerable distances in which the cost per tonne of IOT transported is less than R$ 25. However, scenario 1 was the best alternative, as it admitted a greater distance (290 km) when compared to scenario 2 (170 km). This outcome is due to the higher investment for acquiring the conveyor belt to transport the IOT between the tailings dam and the processing plant.
Lopes (2010) and Ribeiro et al. (2016) compared the use of conveyor belts versus dump trucks in mining transport activities. According to these authors, the advantages regarding the conveyor belts are: low number of operators, since one or two are sufficient to monitor long stretches of belts in operation; transport is not interrupted due to rain and fog, ensuring the system feed rate; and the risk of accidents is reduced due to the low number of people and equipment involved. As for the advantages of using trucks, these authors highlight the following: high operational flexibility since the equipment can be transferred to other operating fronts; little or no equipment assembly is required, therefore the services can start almost immediately; road construction costs for trucks to begin operations are reduced; and activities are not interrupted when a single transport unit is removed for maintenance services.
Nehring et al. (2018) pointed out similar aspects when conducting a comparative study between the use of trucks or conveyor belts for ore transportation. These authors affirmed that the replacement of trucks by conveyor belts has undeniable environmental advantages; however, for this option to be economically viable, it is necessary to consider that: (a) the volume of materials transported must justify the high initial investment in the belt acquisition; (b) the project life must be long enough for the initial investment to be paid; and (c) the electricity cost must be analysed against the diesel cost. In a similar study, Ferreira and Leite (2015) also observed that transportation by electric-powered conveyor belts has a lower environmental impact than transportation by diesel-powered trucks since the Brazilian energy matrix has a significant participation of hydroelectric plants.
Therefore, in purely economic terms, the choice of trucking between the dam and the processing plant is the most advantageous scenario in the present study. However, when it comes to environmental and operational benefits, a more thorough analysis should be performed.
Conclusion
The present work carried out the sizing and economic analysis of a reverse logistics system responsible for transporting IOT stored in a dam to the place where it will be used as pavement infrastructure. The sizing of the system was based on the minimisation of logistics costs and was developed through simulation, since the studied system has stochastic variables, and is subject to resource capacity constraints and operational limitations.
The simulation included a comparative analysis between two scenarios: the use of trucks; or conveyor belts for transporting the IOT between the dam and the processing plant. The simulation allowed quantifying the resources needed in each scenario. Finally, these resources were associated with their respective costs for the evaluation of the economic feasibility of the proposed logistics system.
The main contribution of the present work is this new reuse methodology, which includes the reverse logistics system itself and the simulation for its optimisation, to encourage the reuse of IOT on a mass scale. Although the equipment and techniques adopted are (purposefully) not new, their layout, sequence of use, and sizing method were entirely developed by the authors.
The simulations resulted in distances between the processing plant and the road construction site greater than 150 km that still maintained the cost per tonne of transported IOT below the cost of conventional material (R$ 25). Scenario 1, which opted for trucking in the first route, was the best alternative, allowing a transport distance of 290 km within the established price limit. Scenario 2, which assumes the use of conveyor belts for the initial route, was deemed feasible up to 170 km.
In conclusion, this preliminary analysis resulted in economic feasibility for the logistics systems of both scenarios. Since the equipment adopted is commonly found in mining sites around the world, as well as the specified reuse (road construction), this methodology can be applied to encourage large-scale reuse of IOT in Brazil and worldwide. Environmental and social implications include the reduction of the IOT stored in dams, as well as the risks involving these structures.
One of the limitations of the present methodology is that some infrastructure costs, such as building the roads for truck flow, were not considered. Additionally, the cost of natural aggregate, R$25, was based on a regional average and does not consider local pricing, which can vary slightly.
The present study merely found the greatest distance so that the cost of transport of the IOT was lower than the acquisition cost of the natural aggregate. However, it is known that the reuse of IOT promotes several other advantages for companies, such as reduced dam management and monitoring expenses, minimisation of the risk of accidents and improved corporate image.
Footnotes
Acknowledgements
The authors are grateful for the infrastructure and collaboration of the Research Group on Solid Wastes – RECICLOS – the Brazilian National Council for Scientific and Technological Development.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: The authors are grateful to Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, Minas Gerais State Agency for Research and Development – FAPEMIG and the Brazilian National Council for Scientific and Technological Development for providing financial support.
