Academia.eduAcademia.edu

Figure 1 – uploaded by Felipe Contreras

See full PDFdownloadDownload figure
Parameters of the models fit for batch tests (triplicate). Table 1 conditioned for flotation. Otherwise, the Rectangular model can be shifted on the k axis to separate the floatable valuable from the non-floatable valuable mineral, but in this case the number of parameters becomes similar to that of the Gamma and Weibull functions. Three replicates were carried out obtaining a small variability in the results.

Table 1 Parameters of the models fit for batch tests (triplicate). Table 1 conditioned for flotation. Otherwise, the Rectangular model can be shifted on the k axis to separate the floatable valuable from the non-floatable valuable mineral, but in this case the number of parameters becomes similar to that of the Gamma and Weibull functions. Three replicates were carried out obtaining a small variability in the results.

Related Figures (10)

Fig. 1. Models recovery prediction (replicate 1). 4.2. Plant tests The flotation rate distribution should only consider the valuable mineral which can be floated (k > 0), while the maximum recovery, R,., represents the fraction of valuable mineral which is actually Fig. 1 shows that a single parameter model, such as the Rectan- gular model, can be as good as the Gamma or Weibull distribution in terms of the model fit. However, it fails in terms of identification of the maximum recovery, because it became interrelated with the presence of a mineral fraction with a zero flotation rate constant (Kelly and Carlson, 1991). In fact, Table 1 shows R,, values greater than 99% for Rectangular models whereas R,, ~ 95% was observed with all other models.
Fig. 1. Models recovery prediction (replicate 1). 4.2. Plant tests The flotation rate distribution should only consider the valuable mineral which can be floated (k > 0), while the maximum recovery, R,., represents the fraction of valuable mineral which is actually Fig. 1 shows that a single parameter model, such as the Rectan- gular model, can be as good as the Gamma or Weibull distribution in terms of the model fit. However, it fails in terms of identification of the maximum recovery, because it became interrelated with the presence of a mineral fraction with a zero flotation rate constant (Kelly and Carlson, 1991). In fact, Table 1 shows R,, values greater than 99% for Rectangular models whereas R,, ~ 95% was observed with all other models.
Fig. 2. Rate constant distribution F(k) in Batch test (replicate 1).
Fig. 2. Rate constant distribution F(k) in Batch test (replicate 1).
Gamma distribution parameters (plant case 1).
Gamma distribution parameters (plant case 1).
Fig. 5. Flotation rate constant distribution (plant case 1).
Fig. 5. Flotation rate constant distribution (plant case 1).
Fig. 3. Normalized mineral tracer concentration (plant case 1).
Fig. 3. Normalized mineral tracer concentration (plant case 1).
Fig. 4. Model fit of floatable tracer concentration (plant case 1).
Fig. 4. Model fit of floatable tracer concentration (plant case 1).
Fig. 8. Flotation rate constant distribution (plant case 2). Gamma distribution parameters (plant case 2).
Fig. 8. Flotation rate constant distribution (plant case 2). Gamma distribution parameters (plant case 2).
5. Conclusions Fig. 7. Model fit of floatable tracer concentration (plant case 2).
5. Conclusions Fig. 7. Model fit of floatable tracer concentration (plant case 2).
Related topics:
Chemical EngineeringResidence Time DistributionGamma Function
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved