Improving the rotary kiln-electric furnace process for ferronickel production: Data analytics-based assessment of dust insufflation into the rotary kiln flame

Abstract

Ferronickel contains 60–80 wt% Fe and 40–20 wt% Ni and is a feedstock for manufacturing stainless steel and other ferrous alloys. The primary pyrometallurgical route to produce ferronickel from laterite nickel ores is the Rotary Kiln-Electric Furnace (RKEF) process. In the RKEF process, minerals undergo calcination and partial reduction in a rotary kiln. Large amounts of mineral dust leave the kiln entrained in flue gases. Dust insufflation is a potential solution in which dust particles enter directly into the burner flame. The aim is that the insufflated dust softens, agglomerates, and finally joins the minerals stream that goes into the electric furnace. We applied data analytics techniques to a 1-year operation database to assess the operation before and after dust insufflation in an industrial rotary kiln furnace. Bootstrapping revealed statistically significant differences in the operation with and without dust insufflation. Principal Component Analysis (PCA) and k-means clustering were used as exploratory techniques. PCA showed subsets of variables that significantly influence the dataset variance, and k-means allowed distinguishing operation conditions for dust insufflation with encouraging results for the calcine-to-fresh mineral and the natural gas-to-calcine ratios. These results pave the way towards successfully implementing dust insufflation in the RKEF process.

Keywords

• Nickel laterite
• Calcination
• Reduction
• Agglomeration
• Machine learning
• Data handling