The Research of Fault Diagnosis in Aluminum Electrolysis Based on Rough Set
Jiejia Li, LI Shi-tao, Zhichao Fang · 2008
This paper combines rough set and genetic algorithm with fuzzy theory to diagnose faults in aluminum electrolysis to save energy. Firstly the author gets the simplest decision table by using the rough set to reduce the initial decision table which is made up of the original data. Because of one of important part in rough set being the reduction of condition attribute so a satisfied result can be got by using GA to reduce the condition attribute. And according to the simplest decision table, faults are diagnosed by fuzzy theory. Also, the method of original data pretreatment by rough set has simplified the fuzzy rules, decreased computation and diagnosis time, so the diagnosis efficiency, reliability and precision are obviously improved. The simulation has proved that the method can forecast and diagnose faults actually in the aluminum electrolysis to product aluminum safely and decrease energy consumption.