banner

Industrial companies use a very little amount of real-time data to drive their efficiency. The biggest challenge is to make a fast-track delivery of the data with proper visualization to understand the decision which is the best to come with now. NTWIST has its own decision for this. But let’s recognize the problem first.

Achieving goals in mining

Mining companies are working towards several goals, such as, reducing costs to focus on profitability, improving performance of plants and improving their environmental footprint. 

The question is how can they achieve all these goals? Using data analytics, companies have the ability to analyze plant operations, process setpoints and parameters to better understand how to solve the problems. The first step is for companies to understand their process fully and recognize what the company is missing in value. For instance, plants are experiencing a lot of shutdowns and therefore resources are not being used to their maximum potential. However, they are not able to explain why these shutdowns are happening.  

A business process management tool can be used to help with automating business workflows with the goal of improving performance by minimizing errors and inefficiencies. For example, the nDatum tool created by NTWIST focuses on extracting actionable information from historical data in order to improve performance.

One parameter that has been widely investigated to improve the performance of a plant is throughput. When maximizing the throughput of a plant, it is important to investigate other metallurgical KPIs. For instance, increasing the throughput of a plant can cause reduced ore recoveries and therefore deterring operations from increasing throughput. Historical data should be analyzed to identify if this relationship between recoveries and throughput exists. 

Once the company has recognized that there is potential to increase the throughput of a plant, set inputs need to be identified so we are still maintaining our targets for other metallurgical KPIs. 

Another way to improve the performance of a plant is helping plant engineers and metallurgists make decisions more quickly and accurately. Using technologies related to artificial intelligence, mining companies are able to use their historical data to identify patterns and create insightful decisions. Having inputs that are more reactive to a plant’s data allows for a plant to perform more efficiently with less human error. 

Using the setpoints recommended from operations, it is safe to assume that these inputs are never optimal due to there being so many moving parts in a plant and therefore makes it difficult to predict what the optimal throughput should be. For example, the nDatum tool uses historical data to find the most favorable regimes of operations in order to increase the target KPI without compromising other metallurgical KPIs. 

Discovering missed opportunities to achieve higher recoveries

Looking at a histogram of Leach/CIP Recovery and the number of shifts, there are several shifts when the recovery drops below 90%. There are also more than hundreds of shifts where the recovery is above 90%. This means that there are missed opportunities to achieve higher recoveries that can be investigated using process management tools. 

Eliminating these periods of low recoveries will increase our average recovery by up to 4% which is equivalent to tens of millions a year in additional revenue.  The solution would be to find a number of potential influencer variables will be identified and setpoints will be identified to improve recoveries. An example of an influencing variable is grind size. The effect of grind size on recovery showed that tightening the grind size to 2% will improve the recovery by 0.5 to 1%, equivalent to $2 to $4 million per year. 

Due to the plant being such a complex system, it is almost impossible for humans to select setpoints that will maximize a plant’s performance. As AI-related technologies are slowly being integrated into the mining industry, plant operators and metallurgists are able to make decisions quicker and more accurately. This allows for less human error and an increase in the performance of plants while reducing the environmental footprint. 

9650 20 Ave NW, Edmonton, AB, T6N 1G1, Canada

© 2022. Ntwist Inc.