Freeport-McMoRan has been granted a patent for a method that uses historical and future assumption data to optimize metal production in mining operations. The method involves training a predictive model using historical data, adding future assumption data, comparing forecast data to actual data, determining deviations, and adjusting parameters to optimize metal production. GlobalData’s report on Freeport-McMoRan gives a 360-degree view of the company including its patenting strategy. Buy the report here.
According to GlobalData’s company profile on Freeport-McMoRan, Extraction by leaching was a key innovation area identified from patents. Freeport-McMoRan's grant share as of June 2023 was 1%. Grant share is based on the ratio of number of grants to total number of patents.
Optimizing metal production through predictive modeling and parameter adjustment
A recently granted patent (Publication Number: US11681959B1) describes a method and system for optimizing metal production in mining operations. The method involves acquiring mine plan data from sensors and conducting leaching operations to obtain historical data. A trained predictive model is then used, which incorporates the historical data, to add future assumption data based on the mine plan. The forecast data for a mining production target is compared to the actual data, and deviations between the two are determined. Each of a plurality of parameters from the forecast data is changed to the actual data to determine their contribution to the deviations. Finally, the leaching operations are adjusted based on these parameters to optimize metal production.
The historical data used in the method can include various types of data such as ore map data, mineralogy data, irrigation data, raffinate data, heat data, lift height data, geographic data on ore placement, blower data, PLS (pregnant leach solution) data, stockpile data, or section mapping data. The mineralogy data, for example, can include information on the percentage of copper in a sample, average acid-soluble copper grade, quick leach test percentage, ore size, or tons. The irrigation data can include details on raffinate application rate, area under irrigation, or days under leach.
The method also involves running a forecast engine for the plurality of parameters to obtain the forecast data for the mining production target. Additionally, a predictive model is trained using the historical data to create the trained predictive model. The method allows for overriding the adding and receiving of historical data by inputting different data, including economic and accounting data. The historical data can be cleaned or removed if it is considered bad data.
The patent also describes the use of a machine learning model to create a column test predictive model based on data from a column test. The column test provides output data such as days under leach of a mineral, percentage of each mineral reacted, or amount of acid consumed. Input data from the column test can include raffinate Fe, raffinate acid, leach additive, leach catalyst, raffinate Cu, days under leach, XCu, QLT, application rate, or cure acid.
Overall, this patented method and system provide a comprehensive approach to optimizing metal production in mining operations by utilizing historical data, predictive modeling, and parameter adjustments based on deviations between forecast and actual data.