Made up of many components and complex interfaces, there is often reluctance among diverse stakeholders to cooperate to a level necessary to find the best overall outcome. As a result, supply chains can present a multifaceted problem of different objectives for different users.
The combination of competitive forces, a desire for autonomy and legacy systems which have grown out of a short-term and ad hoc approach to expansion, results in the development of complicated operations and redundant infrastructure.
Improvement in operational efficiency is a current mandate for resources companies around the globe. Routinely, the first step for improvement in most legacy supply chain infrastructure is simply to improve communication and coordination of planning.
However important these initiatives, they are an endgame of incremental improvements for owners and operators. The bigger gains are to be made much earlier in the process, in properly optimising the supply chain configuration for new developments or major expansions early within the project lifecycle.
Inefficient or underperforming supply chain solutions locked between concept and feasibility stages can lead to a number of undesirable outcomes such as:
Currently, there is no regulatory requirement to demonstrate a rigorous and quantifiable approach to optimising the configuration and operating plan of the supply chain to maximise economic benefit. Project fundamentals are commonly shaped at concept stage, based purely on past experience and rule of thumb. As a consequence, it is likely that the set of options to be considered in subsequent project stages will not contain any near-optimal supply chain configurations.
“In supply chain projects with a heavy haul rail component, we see this when intuition, or preconceived notions that longer trains are more efficient, form the basis for a choice of what is known as 'consist' length,” says Colin Eustace, technical leader for operational optimisation, Aurecon.
“Such a choice is often made before the development of any rail infrastructure options and such an arbitrary choice of length can shape all of the options developed, right from concept study inceptions. This approach may lead to exclusion of future near optimal solutions from consideration".
Fixing the target throughput at the concept stage is common and often prevents the economic optimisation process from considering optimal solutions. Optimisation of transport supply chain and mine planning should be integrated; so that the optimum mine production profile can be determined, along with the transport supply chain configuration. In some cases, it may be beneficial to reduce mine production by a small percentage (e.g. targeting a higher grade product) if it means avoiding a step change in supply chain capital cost.
“Using intuition to shape projects at concept stage is understandable to an extent. The decision space is vast, with many variables involved in complex relationships,” comments Colin. “Even for one given solution, it is difficult to forecast system performance in terms of throughput and cycle times.”
Applying intuition and preconceived ideas provides a means of narrowing down the decision space to a manageable size, but usually at the expense of even considering options that may represent significantly better value.
Using the illustration of a greenfield bulk supply chain involving heavy haul rail and a bulk port, this discussion provides an overview of a framework to develop a comprehensive quantitative evaluation of the decision space for major supply chain projects. It also includes an argument demonstrating that optimality in terms of economic benefit should be a regulatory requirement for major projects which exploit a nation’s resources and significantly impact the environment.
The best sustainable development outcome considers more than just the economics of the operation. Environmental and societal outcomes, as well as capacity, reliability and time–to–market all play an important part. These objectives can all be quantified and considered in numerical optimisation. The discussion uses the term numerical optimisation to differentiate the science of mathematical optimisation from a structured process of intuitive scenario development, comparison and ranking, and concentrates on optimisation of the economics of the supply chain. With appropriate legislative safeguards in place, good economic project outcomes generally lead to good environmental and societal outcomes by increasing the availability of funding for environmental and social programmes, and increasing government tax revenue.
From a supply chain economics perspective, the best outcome is a supply chain that achieves the lowest breakeven transport and handling cost. This implies a minimum combination of initial capital expenditure and operating expenses over a particular investment horizon and with an assumed discount rate. By breaking potential supply chain alternatives into capital and operating expenses (fixed and variable), each option can be compared on a like for like present value (PV) or breakeven transport cost basis. Further, by parameterising the variables that make up a transport solution, such as spacing of rail passing loops, conveyor rates, rolling stock characteristics and train consist size, it is possible to evaluate and compare a full spectrum of supply chain alternatives. The results can be interesting and often suggest different configurations to intuitive solutions and sometimes with significantly better outcomes.
For a given set of cost parameters, it is possible to demonstrate that a supply chain is an optimised design.
Surprisingly, very few resource operations take such comprehensive quantitative measures to ensure that a proposed new supply chain or upgrade is economically optimised.
Supply chains are typically complicated configurations of legacy infrastructure, multiple users and unique attributes. At first inspection, such an operation may not seem an appropriate candidate for numerical optimisation at a whole of supply chain level. Nevertheless, concept level optimisation is generally effective in establishing the right supply chain operation for further refinement through to detailed design and construction. This is best demonstrated using an example of a proof of concept greenfield bulk supply chain development, although the approach can be equally effective for analysing brownfield operations versus their theoretical optimised state. In practice, supply chain solutions are invariably complicated by topography, large bridge structures, environmental considerations and other idiosyncrasies. The influence of unique attributes of a particular supply chain can normally be accounted for within the described optimisation process.
As system capacity is increased, the break-even rail haulage costs for a 300 kilometre distance between mine loadout and dump station can be plotted to reveal cost benefits across different rail configurations (See Figure 1). As expected, break-even haulage cost per tonne decreases with larger throughput and increases with haulage distance. Higher capital expense – lower operating expense solutions are also preferred at higher tonnage levels. The lowest curve at a particular tonnage indicates the ideal type of rail operations at a particular throughput level. As throughput is increased, the transition to electrification and then duplication of the initial single track diesel operation is shown.
Clear relationships emerge through the comparison of the minimum break-even transport costs across the spectrum of supply chain configurations.
As supply chain throughput is increased, the minimum break-even haulage costs can be calculated for different length supply chains from mine loadout to export terminal and for electrification and duplication options for each length (See Figure 2). The red line shows the point at which electrification becomes a more cost-effective option over a diesel single track with passing loops operation. The blue line shows the point at which duplication provides optimised project value. The trigger for duplication occurs at much lower throughput levels if the option of electrification is not available.
The same proof of concept example is used to focus on optimisation of consist length (see Figure 3). The option of electrification is not considered in this case to provide clearer consist length outcomes. The red line indicates the point at which double tracking and using short train lengths becomes a more cost-effective option over single track with passing loops and long train consists. Preferred train consist length is mapped against throughput and haulage distance.
The example supports the industry rule of thumb that many practitioners are familiar with in that longer haulage distances and higher throughputs generally suggest longer train consist lengths. However, the quantitative analysis also suggests that double track infrastructure with relative short consists and fast turnaround times, would be a far more cost-effective supply chain solution to many of the existing bulk supply chain systems in markets such as Australia (i.e. the Pilbara), provided a long term investment view is taken.
A simple concept level model of a generic greenfield bulk resources supply chain forms the basis of the described proof of concept example. The configuration of the model has many assumptions concerning operational details, capital and operating costs, which may vary significantly from supply chain to supply chain. Although not detailed here, the assumptions (i.e. cost of rail infrastructure, terminal equipment, rolling stock, diesel, electricity, etc.) used are broadly similar to most iron ore or coal supply chain operations. While many cost parameters used in the proof of concept model influence the relationships developed, the relativities between alternative configurations are fairly constant. As a result, relationships between alternative configurations do not change significantly within a reasonable cost range for the majority of parameters and results are likely to be indicative of many supply chain operations, whether in Australia, South Africa, Sub-Saharan Africa, etc.
However, given a comparison of relative costs drives the triggers for electrification and duplication, the best throughput to transition from one type of operation to another may vary significantly for particular systems. The intention of the data shown is merely to demonstrate that cut offs between alternative configurations do exist and can be quantified. Emphasis should be on the general shape of the relationship between alternative operations rather than the exact location of trigger points.
A heavy haul rail and bulk material handling terminal to ship operation is configured to meet a range of throughput requirements and haulage distances. Rail infrastructure and terminal configuration details are parameterised so that changes to target throughput and other scenario parameters will trigger a reconfiguration of the infrastructure, rolling stock, required rail operations and associated capital and operating costs. By taking a parameterised approach to the economic modelling, a comprehensive spectrum of alternative concept level rail and export terminal configurations can be evaluated. A minimum breakeven cost per tonne is then calculated for each throughput haulage distance combination, considering a spectrum of options for parameters such as: consist length, terminal configuration, and single track vs. duplication in each case.
Numerical optimisation might be achievable for a new, idealistic, vertically integrated supply chain, but how does it apply to legacy systems with multiple users?
There is actually very little difference between applying whole of supply chain numerical optimisation to brownfield expansions and new greenfield developments. The only difference is the subtraction of the value of the existing useful infrastructure and rolling stock from the capital cost estimates for the required infrastructure of the optimised scenario. The capital cost difference influences the outcome of the numerical optimisation.
Brownfield expansions are typically focused on the lowest capital cost short term solution that meets the target capacity for an expansion. The best upgrade option should be identified by comparing the existing system and current capacity with an optimised supply chain configuration at target capacity. Consideration should also be given to whether short term decisions align with the optimised configuration for further expansions. Consider a system where the throughput figure which triggers duplication of the mainline is just around the corner. Once the mainline is duplicated, the optimum consist length may be much shorter than the existing length, taking advantage of additional mainline capacity to achieve faster cycle times with shorter trains (reduced loading and unloading times). The total required rolling stock fleet may be reduced as a result. Investing in longer train consists in the short term might be a poor long term decision, as part of the rolling stock fleet adopted for the short term expansion would become idle once the mainline is duplicated.
If it is accepted that the example given suggests supply chain PV can be provably optimised, the ensuing question is: “How could this be implemented in practice?” Potentially significant improvements are only possible if the initiative to do the analysis is freely taken or the analysis is a regulatory requirement of undertaking a project.
“Ultimate responsibility for ensuring that a resources project has been optimised over the long term rests with the government, although suitable industry incentives are required to facilitate sharing of relevant information and enable alignment of long-term planning amongst supply chain stakeholders,” says Colin.
In the same way an environmental impact statement is a mandatory requirement for resources projects, assessment of fit for purpose design for logistics infrastructure should be a regulatory requirement.
A project optimisation statement could be used to demonstrate that a proposed project had been optimised to maximise economic benefit. The statement could be evaluated against a framework to identify whether sufficient analysis to demonstrate an optimised and robust design and operating plan has been developed and challenged through peer review by a third party. This process would be relatively difficult to implement for a number of reasons, not least of which is the ability to numerically optimise resource supply chains is not widely available.
However, closer scrutiny of proposed supply chains from the perspective of all potential users of the infrastructure has the potential to save billions of dollars in capital and operating expenses. Resulting savings could have significant impact on profit and therefore government revenue and funds to improve environmental and societal outcomes.