Use Of Published Data In The Development Of Hydrometallurgical Flow Sheet For Gold Using Decision-Support Tools

The design of a hydrometallurgical process starts with the analysis of the raw materials to be treated, for example mineralogy, chemistry, deposit size are analysed. Based on this information suitable unit processes for pre-treatment, leaching, solution purification and product recovery are selected. Typically, process design for the new ore deposit relies on laboratory tests such as pre-treatment and leaching experiments. The experimental work is the most time consuming and expensive part of the process design. Reducing the number of experiments would shorten the time of designing a new flowsheet or development of new process technology. A large amount of information is available about the hydrometallurgical experiments and the amount is increasing continuously. For example, word search using ?hydrometallurgy? from Scopus database results in over 4 000 publications with the word in title, abstract or keywords. This number consists of 2 500 articles, 900 conference papers, and 90 reviews since 1972. The number of hydrometallurgical publications per year has increased from 13 in 1972 to 150-200 in the last few years. The objective of this study was to analyze how detailed data of experiments about the gold extraction is published and how that data could be exploited more effectively when designing hydrometallurgical processing for new ore. The leading idea is that similar ores should have the similar response in leaching. If the published information was used more efficiently, the amount of experiments would be reduced. More than 200 cases of experimental data of gold leaching have been analyzed for this study. There are over 30 factors that could be used in comparison when selecting process alternatives, but challenging fact is that not all these factors are documented in all publications. Results show that the usefulness of the published data in the development of hydrometallurgical flow sheet for gold using decision-support tools is not that straightforward. The quality of the published information about gold extraction varied between the publications and the descriptions of the raw materials, experimental arrangements and results were often missing details. Methods to normalize the data for description of the cases and a system to describe different features unambiguously are needed. Keywords: refractory gold, host mineral, leaching, data analysis, case-based reasoning