The production of manganese alloys is critical in steel production, batteries, and various industrial applications. For smelting operations that produce manganese alloys, the price and cost of the materials used to create alloys are constantly changing. Prices, availability, and chemical composition fluctuate, creating a challenge to profitability. Generating new recipes can be a challenge. Ensuring those recipes are also the lowest cost recipes is even more difficult. This is where mathematical optimization can be a game-changer for smelting optimization.
To manage production, smelting operations often rely on historical plans, manual adjustments, and the use of the same recipes repeatedly. Despite relative price spikes, many operations stick to last month’s recipe in order to save time instead of adjusting away from relatively expensive ingredients or towards relatively cheaper ingredients. As a result, they can leave potential for money-saving solutions, which makes it a challenge to increase profitability. To save time and continue production, this approach leaves mining operations in a state of survival mode.
This is where mathematical optimization can come into play, allowing metallurgists to take a more autonomous and optimized approach to recipe generation, which brings an operation from surviving to thriving. Smelters often swing between profitability and loss. One of the biggest variables in the expenses column is the cost of ingredients. A 2%-5% reduction in costs can double operating income.
The Benefits of Mathematical Optimization in a Volatile Market
With the standard practice of copying last month’s recipes and making minor tweaks, you are leaving money on the table. Prices of manganese ore and other additives can fluctuate from week to week, and manual adjustments require too much time. Optimization models, on the other hand, can continuously analyze market prices and availability for all raw materials to calculate the most cost-effective blend that meets product specifications at any given time. This leaves you with opportunities to utilize your current inventory, complementing it with available ingredients, and using less of the expensive materials when prices rise and more of them when prices fall.
Engineers can benefit from utilizing these optimization models to adjust recipes more efficiently, ensuring they always use the correct proportion of available ingredients while maintaining consistent quality standards and drastically reducing manual rework and off-spec material.
Optimized operations can also reap the benefits of evaluating new raw materials sources or alternative ingredients that suddenly become available or cheaper. Complex interactions with materials are too hard to calculate manually, on the fly, but optimization models can make this challenging and sometimes impossible task simpler.
Solve Operational Bottlenecks and Improve Inventory Management
In addition to an optimized recipe generator, optimization models can resolve operational bottlenecks in real-time with current data and ensure maximum and efficient material flow through the operations process, regardless of market conditions and ore input.
Price volatility often impacts demand and desired output. Optimization models can adjust mine-to-plant schedules based on real-time market conditions, helping you decide when to ramp up production to capitalize on demand or slow down production to avoid oversupply.
With optimization models, you can also determine the ideal inventory levels for different materials, ensuring a proper flow and stock of materials and reacting to price forecasts and supply chain stability.
PhD Corner- Expert Advice
Get answers from our PhD Expert: Joshua Woodruff
When manganese prices spike, how quickly can an optimization model actually recommend a “new best recipe” or schedule adjustment, compared to a manual process?
An optimization model can generate an entire new monthly plan in minutes. It can generate a new recipe in seconds.
Can you give an example of how the optimization considers not just the current price of manganese ore, but also the fluctuating prices of other key inputs like quartz and dolomite to find the most profitable blend?
There are multiple grades of ore. Depending on their relative prices, it may be better to generate recipes that use high quantities SiO2 and Al2O3 in the slag or using more CaO and and MgO in order to adjust the ratio of MnO in the slag to Mn in the alloy.
Beyond just cost, what other factors does a mathematical optimization model typically balance when trying to create an ‘optimal’ schedule or recipe for a manganese operation facing price volatility?
An optimization model doesn’t just look at costs. It looks at limits to energy and ingredient availability. There may be a very low cost recipe, but if there aren’t enough ingredients to meet all of the demand, optimization needs to create more recipes that use the ingredients that are available.
Are costs the only thing optimization considers?
In the short term, costs are the main driver. This is to ensure the smelter is able to meet all customer commitments. But when thinking longer term, the optimization maximizes profitability. It can recommend production quantities that leverage available ingredients. Above a certain quantity, and the cheap ingredients may be limited, causing the costs per ton to increase with volume, decreasing profit. The highest profits don’t necessarily come from producing the most.
Conclusion
Price volatility can present challenges in recipe generation and slow down decisions if done manually. Mathematical optimization presents an opportunity for mining operations with machine learning models that can help alloy smelters grow by embracing the constant market price changes and maximizing the use of their inventory.
Mathematical optimization is here to help you have a competitive advantage and allow manganese operations to adapt swiftly. Helping you make data-driven decisions and ensure profitability even in the most unpredictable markets.
Are you curious how an optimization model can tackle the complex challenges of your mine? Schedule a quick chat with our PhD experts.
Modaai has Operations Research PhDs with over 25 years of experience in modeling and tuning. We can tackle your toughest problems. Whether you want us to take over your optimization project completely or supplement your team with training, model design, or model tuning, we can provide the support you need.