escondida prospects

Maximising value: four technologies that return dollars, time, and tonnes

Laura Tyler

Laura Tyler

Chief Technical Officer

BHP’s ability to create value is underpinned by our performance. We constantly look for ways to make our operations more efficient.

For more than a decade, an enhanced approach to technology has unlocked significant performance at BHP through innovations such as remote operations centres, fleet management systems and automated equipment, like trucks and drills.

Now we are taking operational technology to the next level using data, machine learning and decision automation to solve business problems, improve safety and increase productivity.

It is not necessarily the most complex technology that creates the most value and we have found that some of the simplest solutions have been the most effective in enhancing our daily operations.

Here are four ways we have used technology to increase returns, time and tonnes applied to our operations today:

Reduced product variability provided a US$10 million revenue uplift 

In iron ore mining, grade variability is a key factor in product price, so reducing variability adds to the value of our products in the market. Given the volume of tonnes we move, just a small reduction in variability makes a big difference both to the value of our products and to the satisfaction of our customers. 

At our Jimblebar iron ore operation, we developed an in-house machine-learning model to reduce iron ore grade variability across the value chain. We pull data sources from ore movements across Western Australia’s Pilbara region to map the iron ore grade from the mine to the product shipped at Port Hedland. The model predicts grade biases daily and updates mine plans and targets accordingly. Our teams then develop the schedules to optimise the grade requirements to customers’ needs.

With the application of this modelling, there has been a US$10 million revenue uplift at Jimblebar in the past 24 months. Replication at other Western Australia Iron Ore (WAIO) sites is estimated to double the revenue benefit to upwards of US$20 million per year and provide improved service to our customers.

Optimised schedules reduced rail grinding downtime by 45 per cent

Rail grinding to maintain optimum rail performance and rail life is the cornerstone of virtually every railroad asset integrity maintenance program. The task is complex. Grind too much and you consume too much track time; grind too little and you risk defects on the track creating a potential safety or production issue.

Many factors come into play when this work in planned: the size and complexity of the WAIO network, grinding machine availability, and human constraints such as crew size, shutdown work and even extreme weather. It is no surprise grinding takes the WAIO railway offline more often than any other activity and accounts for around 30 per cent of rail downtime. 

Thanks to machine learning and schedule optimisation techniques, we have improved the grinding schedule and reduced the required annual grinding hours by 45 per cent. As an additional bonus the compliance to schedule has increased to 94 per cent. Both are fantastic results that keep our product moving and improve safety on the ground through more and better-planned work.

Fast-tracked material movement increases outflow capacity by 1.4 million tonnes per year

The material handling network at Port Hedland is complex. It consists of five car dumpers, eight stackers, five reclaimers, three lump rescreening plants and eight ship loaders, all connected by around 100 conveyor routes.

Selecting the fastest route to send the right material to the right ship is a challenge and technology is proving a game-changer.

We now use algorithms to identify high and low performing car dumper routes, then select those that reduce dump times and positively impact vessel line-up. These algorithms give us the fastest pathway to move material without guesswork.

We have measured an uplift in outflow capacity of 1.4 million tonnes per year with very limited capital investment. This results in significant additional value sustainably and reliably delivered.

Better customer schedules deliver over US$1 million additional revenue per month 

In most situations, any time we can save in any process will contribute to efficiency and ultimately revenue.

At our copper operations in Chile, it used to take half a day to generate a monthly customer shipping schedule for the approximately 40 individual copper concentrate customer sales that we manage each month from our Escondida mine. Until recently, this complex task was performed manually through spreadsheets, with variables such as product quality and changing commercial terms making the task of scheduling very complex and time-consuming.

Today we use an in-house machine-learning tool called Trident that reduces the planning window to just 30 minutes each month. Trident uses data analytics to offer scenarios and output projections to allow us to dynamically adjust or flex tolerances in each customer delivery. We now continuously optimise schedules in response to mine plan changes, vessel arrival times and/or customer requests to deliver the best commercial outcome every time.

Simply put, Trident maximises value and minimises risk – we now allocate better quality and better aligned products to our customers’ needs. It minimises operational disruptions because we can quickly solve for variations in the mine plan with less impact to our schedule and customers.

Since we implemented Trident in November 2020, we have recouped an additional US$1 million each month for copper concentrates produced at Escondida, reduced penalty discounts and improved payables, plus customer service levels have risen significantly. We now plan to implement Trident across the broader concentrate business, including at Spence, and we are considering how to maximise its full potential. We see that Trident can deliver volume upside for many of our assets as forward logistics schedules are paired with production schedules to develop and select optimised value chain scenarios. 

The potential application of many of these technologies is vast, and these are just a few examples of the exciting work underway to unlock further value and deliver the next step change in safety and productivity.