In the scope of this project, the specific industrial target is to improve the physical and technological properties of ceramic pastes produced from two different types of raw material, feldspathic ores and raw clays. In order to accomplish the objective, the chemical/mineralogical composition of the raw-materials should satisfy some specifications. Froth flotation is the main mineral processing unit operation able to improve and/or control the raw material specifications.
Regarding feldspathic ores, a great number of flotation tests (bench scale and pilot plant scale) were carried out with very good results in what concerns the mica (and other heavy dark colour minerals) removal and the quartz/feldspar ratio control of the final product. Dry high intensity magnetic separation was also applied to one feldspathic ore sub-type with excellent results. Regarding the raw clays, iron grade reduction was succeeded, in one sub-type, by froth flotation. However, iron grade reduction was achieved mainly by size classification using spiral classifiers and hydrocyclones.
To deal with the wild fluctuations of the chemical/mineralogical composition of the raw- materials, the quality/productivity parameters (grade and recovery) of the final product are kept within predefined limits by one controller inspired in fuzzy logic systems. The correspondent WINDOWS 95/98 real-time control application was developed and tested in a fully instrumented (including one on-stream XRF analyser with two sources for light and heavy chemical elements) flotation-based pilot plant, which was completed in the scope of the project. Several papers were already published reporting the results obtained. A collaborator of the project has finished, with success, her Ph.D. thesis dealing with the application/evaluation of real-time fuzzy control to a flotation column.
The controlled blending of several (around 10) different raw clay types is the present practice of one partner to meet the client specifications. An expert experimental approach is used to find out the weight proportions of the raw clays to be blended. Intending to replace this time-consuming approach, the development of a WINDOWS 95/98 software tool to define, automatically, the weight proportions of several different raw and processed clay types in the final blending, satisfying the required ceramic physical/technological properties, was fully accomplished.
Finally, large samples (about 100 kg) of the different processed raw materials were produced in order to perform a final checking ceramic tests by one of the partners (end user). The results confirmed superior properties of the ceramic paste produced from mineral components processed in the scope of the project.