Estimating Residence Time Distributions in Industrial Closed-Circuit Ball Mills

This paper compares two deconvolution methodologies used to estimate residence time distributions (RTD) in industrial closed-circuit ball mills. Parametric and non-parametric techniques were evaluated. Both allowed for direct RTD estimates from inlet outlet tracer measurements the mills, with no need mass balances nor assumptions correct effect of recirculation grinding circuits. Measurements concentrations conducted by radioactive solid tracers on-stream detectors. The parametric was applied assuming N-perfectly-mixed-reactors-in-series model, whereas consisted a constrained least squares estimation subject non-negativity. shapes estimated RTDs consistent between these methodologies, showing mound-shaped all cases. From approach, mixing regimes described 2–4 perfect mixers series observed, which indicated significant differences regarding mixing. mean (τmean) median (τ50) times more when applying deconvolution. approach sensitive noise, disadvantage leading significantly higher than median, less locations. comparisons, strategies proved be applicable led better overall performances τ50 = 1.7–8.3 min, given suitable model structure RTDs.