984 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
type analyses where it is necessary to deal with discrete, dis-
continuous decision variables which cause computational
issues when coupled with rigorous thermophysical property
models. Due to this, it is necessary that PrOMMiS support
a range of thermophysical property models with varying
degrees of rigor, ranging from empirical models with simple
algebraic forms suitable for conceptual design through to
rigorous predictive models which can be used for detailed
simulations and uncertainty quantification studies.
Another challenge facing novel processes in general,
and specifically Rare Earth Elements (REEs)—a key critical
mineral (CM) of interest to the DOE—is a lack of under-
standing of the fundamental thermophysical properties
involved. Whilst a number of predictive models exist for
aqueous and solid phase properties, the necessary data and
parameters for many of the REE species and novel reactants
are poorly characterized. This has so far limited the applica-
tion of these models.
One opportunity to address these challenges in an
expeditious way is the use of existing computational tools.
A number of computational packages exist for modeling
the thermophysical properties of aqueous solutions, such
as OLI (a proprietary tool) and PHREEQC (open-source).
(OLI Systems, Inc., 2024) (United States Geologic Survey,
2021) Whilst these tools provide accurate predictions for
properties of a number of different species of interest,
there are a number of challenges in applying these within
PrOMMiS. Firstly, equation-oriented modeling environ-
ments such as IDAES and the solvers they use rely upon
having access to the first and second partial derivatives of
all equations in the model. These are generally not avail-
able within external packages and therefore precludes them
from being directly incorporated into the models.
One pathway to incorporate these capabilities, how-
ever, is the use of surrogate models. The IDAES-IP provides
a number of tools for developing data-driven surrogate
models where data is generated using an external tool or
package over a certain concentration or operational range
of interest for use within the platform. The PrOMMiS
team is presently working to leverage these external tools to
provide data to train surrogate models for rigorous thermo-
physical models be incorporated into select process flow-
sheets of interest. As robust as these external libraries and
packages are, however, there are still substantial data gaps.
Notably, these tools are often lacking data for key species,
especially the REEs, thus it will be necessary to collaborate
with experimental partners to gather data for many of these
species of interest.
Unit Model Libraries
In order to support he rapid development of models for
novel processes, PrOMMiS is building upon the existing
IDAES-IP model libraries to provide models for a range of
common minerals processing operations. These models are
being developed in a generalized, modular fashion to allow
them to be adapted to a wide range of applications and
used as building blocks for more complex unit operations.
These models are also being developed to support a range of
levels of rigor in order to allow the same models to be used
for conceptual design through to robust optimization and
uncertainty quantification.
Due to the “deploy now” mandate from the Department
of Energy, the initial round of models focused on low-
fidelity models relying primarily on partitioning coefficient
or similar single parameter performance equations with
additional rigor being added where possible based on data
availability. A number of common unit operations for REE
recovery processes were implemented based on the literature
review discussed earlier, supported by data from an FECM
funded pilot study led by the University of Kentucky where
available. (Honaker, et al., 2021) The first round of unit
operations included in the model library include:
• Leaching reactors
• Solvent Extraction
• Selective Precipitation using Oxalic Acid
• Roasting and Calcination
• Solids Grinding
• Clarifiers and Thickeners
• Solar Evaporation Ponds
Additionally, IDAES-IP already includes a number of
other unit operations of interest in CM &REE process-
ing through the WaterTAP and IDAES-CMF model librar-
ies. (National Energy Technology Laboratoratory, 2024)
(National Energy Technology Laboratory, 2024)
These include:
• Nanofiltration and Reverse Osmosis membranes
• Ion Exchange
• Electrodialysis
• Pumps
• Mixers and Separators
Future development of the model libraries will focus on
novel processes such as Membrane Solvent Extraction and
Membrane Distillation, as well as increasing the rigor and
validation of the existing model libraries.
type analyses where it is necessary to deal with discrete, dis-
continuous decision variables which cause computational
issues when coupled with rigorous thermophysical property
models. Due to this, it is necessary that PrOMMiS support
a range of thermophysical property models with varying
degrees of rigor, ranging from empirical models with simple
algebraic forms suitable for conceptual design through to
rigorous predictive models which can be used for detailed
simulations and uncertainty quantification studies.
Another challenge facing novel processes in general,
and specifically Rare Earth Elements (REEs)—a key critical
mineral (CM) of interest to the DOE—is a lack of under-
standing of the fundamental thermophysical properties
involved. Whilst a number of predictive models exist for
aqueous and solid phase properties, the necessary data and
parameters for many of the REE species and novel reactants
are poorly characterized. This has so far limited the applica-
tion of these models.
One opportunity to address these challenges in an
expeditious way is the use of existing computational tools.
A number of computational packages exist for modeling
the thermophysical properties of aqueous solutions, such
as OLI (a proprietary tool) and PHREEQC (open-source).
(OLI Systems, Inc., 2024) (United States Geologic Survey,
2021) Whilst these tools provide accurate predictions for
properties of a number of different species of interest,
there are a number of challenges in applying these within
PrOMMiS. Firstly, equation-oriented modeling environ-
ments such as IDAES and the solvers they use rely upon
having access to the first and second partial derivatives of
all equations in the model. These are generally not avail-
able within external packages and therefore precludes them
from being directly incorporated into the models.
One pathway to incorporate these capabilities, how-
ever, is the use of surrogate models. The IDAES-IP provides
a number of tools for developing data-driven surrogate
models where data is generated using an external tool or
package over a certain concentration or operational range
of interest for use within the platform. The PrOMMiS
team is presently working to leverage these external tools to
provide data to train surrogate models for rigorous thermo-
physical models be incorporated into select process flow-
sheets of interest. As robust as these external libraries and
packages are, however, there are still substantial data gaps.
Notably, these tools are often lacking data for key species,
especially the REEs, thus it will be necessary to collaborate
with experimental partners to gather data for many of these
species of interest.
Unit Model Libraries
In order to support he rapid development of models for
novel processes, PrOMMiS is building upon the existing
IDAES-IP model libraries to provide models for a range of
common minerals processing operations. These models are
being developed in a generalized, modular fashion to allow
them to be adapted to a wide range of applications and
used as building blocks for more complex unit operations.
These models are also being developed to support a range of
levels of rigor in order to allow the same models to be used
for conceptual design through to robust optimization and
uncertainty quantification.
Due to the “deploy now” mandate from the Department
of Energy, the initial round of models focused on low-
fidelity models relying primarily on partitioning coefficient
or similar single parameter performance equations with
additional rigor being added where possible based on data
availability. A number of common unit operations for REE
recovery processes were implemented based on the literature
review discussed earlier, supported by data from an FECM
funded pilot study led by the University of Kentucky where
available. (Honaker, et al., 2021) The first round of unit
operations included in the model library include:
• Leaching reactors
• Solvent Extraction
• Selective Precipitation using Oxalic Acid
• Roasting and Calcination
• Solids Grinding
• Clarifiers and Thickeners
• Solar Evaporation Ponds
Additionally, IDAES-IP already includes a number of
other unit operations of interest in CM &REE process-
ing through the WaterTAP and IDAES-CMF model librar-
ies. (National Energy Technology Laboratoratory, 2024)
(National Energy Technology Laboratory, 2024)
These include:
• Nanofiltration and Reverse Osmosis membranes
• Ion Exchange
• Electrodialysis
• Pumps
• Mixers and Separators
Future development of the model libraries will focus on
novel processes such as Membrane Solvent Extraction and
Membrane Distillation, as well as increasing the rigor and
validation of the existing model libraries.