1. Designing gold extraction processes: performance study of a case-based reasoning system
- Author
-
Thomas Roth-Berghofer, Maria Leikola, Olof Forsén, Christian Sauer, Lotta Rintala, Jari Aromaa, and Mari Lundström
- Subjects
Engineering ,Decision support system ,Chemistry(all) ,Process (engineering) ,Process design ,02 engineering and technology ,010502 geochemistry & geophysics ,01 natural sciences ,Intelligent-systems ,Flowsheet recommendation ,0202 electrical engineering, electronic engineering, information engineering ,Relevance (information retrieval) ,Case-based reasoning ,Process engineering ,ta215 ,0105 earth and related environmental sciences ,business.industry ,Mechanical Engineering ,Metallurgy ,Intelligent decision support system ,General Chemistry ,Geotechnical Engineering and Engineering Geology ,Knowledge modelling ,Test case ,Knowledge-management ,Control and Systems Engineering ,020201 artificial intelligence & image processing ,business ,Gold extraction - Abstract
This paper presents a method for externalising and formalising knowledge involving the selection of hydrometallurgical process flowsheets for gold extraction from ores. A case-based reasoning (CBR) system was built using an open source software myCBR 3.0. The aim of the systems is to recommend flowsheet alternatives for processing a potential gold ore deposit. Nine attributes: Ore type, Gold ore grade, Gold distribution, Gold grain size, Sulfide present, Arsenic sulfide, Copper sulfide, Iron sulfide and Clay present were modelled and several literature sources of actual gold mines and processes were used for acquiring cases for the system. After preliminary testing, functional evaluation of the built CBR system was carried out by using five real mining projects as test cases. Additionally, human experts in the field of gold hydrometallurgy were interviewed to demonstrate the benefits of the CBR system as it holds no human biases towards any processing techniques. It was found that the suggestions of the CBR system provided useful information and direction for further process design and performed well compared to the interviewed human experts, thus confirming that the system is of practical relevance to the process engineer designing an industrial gold processing plant. The current model was found to be a functioning basis for further development through additional attributes, adjusted attribute weighting and increased number of cases.
- Published
- 2017
- Full Text
- View/download PDF