Numerical Modelling in Predictive Mineral Discovery: Geochemical Models F1-2 pmd Team Thursday 4 th September 2003
Key F1/2 Workflow Modelling mineral deposit geology and fluid processes using equilibrium reactor approach (HCh software) Developing user friendly front (ELF) and back (PIG) ends for HCh, easier problem definition and easier result viewing Developing algorithms to model key processes in ore deposit formation with both academic and industry use in mind Approaching this development work with the computer grid visions of the M-people ( The Matrix?), and the industry use visions of the pmd*crc sponsors
HCh-what? HCh (Yuri Shvarov & Evgeniy Bastrakov) uses gibbs energy minimization to locate the equilibrium point of any system. This is a different approach from log K-type modelling, although the end point should be the same result. Advantages to pmd: Powerful and flexible algorithm generator that can be used to model a much wider range of geological (fluid-rock) scenarios. Well maintained high PT thermo datatset that will soon be online thanks to GA developers. Close collaboration with code developer(s) that provides greater potential in future development directions industry focus!
HCh concept Input A A+B B Control P-T Path Output
HCh Control File The HCh control file uses simple algebraic notation to handle the interaction between the input systems and PT. [*] = [1]+([2]*10^(i-6)) Key workflow problem: Geological/ore deposit process concept Algebraic control algorithm The control file manipulation will become more straightforward in ELF (or daughter of ELF), and all plotting needs will be handled by PIG (due Nov 03).?
Examples from Ernest Henry Biotite halo More pyritic rims Chalcopyrite rich (higher grade) core
EHM Mixing: Actual vs 1D Model Plagioclase Actinolite Calcite Chlorite Biotite K-feldspar Muscovite Gold Chalcopyrite Pyrrhotite Pyrite 100% 80% Much better result if Au with HCOS fluid, Cu with brine, and 10% wallrock as buffer K-altered host volcanic MixingD1b Biotite Chlorite Titanite Magnetite Quartz Actual paragenesis Gold co-precipitates with cpy 60% 40% 20% 0% K-fspar Cpy Muscovite (Gold) Ilm Py Po Rutile Magnetite Quartz 5 6 7 8 9 10 11 12 13 titration step brine HCOS
2D Grid simulations: Assemblage plots 2D grids of 3 component mixing (Fluid-Fluid-Rock) C Increasing rock component Fluid Mixing (X B ) C A B
Pyrite Chalcopyrite XRock HCOS Brine
Control File Algorithms Control file algorithms can become quite complex in order to model key fluidrock interaction concepts This example mixes two fluids in the presence of rock and looks at the passage of outflow fluid into wall rock Once you understand the concepts behind the control file you can conceptualise a large range of ore-forming processes * * * Primary wave * * * T = 450 P = 2500 [*] = [1]+(0.1*[5])+(0.1*[8]) 2 450 o C Cu-Brine General step... T = (i/60)*550+(1-i/60)*450 P = 2500 [*] = ([1]*(1-(1/60)*i))+(([2]/2)*((1/60)*i))+(0.1*[5])+(0.1*[8]) Stop when: i=60 * * * Secondary wave * * * T = 450 P = 2500 [*] = {A}+(0.1*[5])+(0.1*[8]) General step... T = 450 1 P = 2500 [*] = {A}+((0.1*[5])+(0.1*[8])) Stop 550when: o C i=60 Stop Au-HCOS when: N=40 450 o C Hm-bearing volcanics
2D Grid simulations: Outflow models 2D grid of dependent reactors (i.e. fluid originates in mix zone) C C Distance into wall rock A B Fluid Mixing (X B )
Ideal Solid Solutions Cross check prediction vs reality
Alteration patterns Alteration patterns in the outflow zones form the fluid mixing site vary slightly depending on the proportion of brine:hcos fluids that mixed. The most distinctive change in major element mineralogy is associated with the reactions between Muscovite = K-feldspar +- Biotite Use this relationship generate plot for: biotite/(biotite + muscovite + k-feldspar) (volumes) Termed X(Bt*)
X(Bt*) values Modeling is iterative Reality checking! Relating this value to actual distance is dependant on the amount of fluid the rock sees (time integrated fluid flux) Need feedback from fluid-flow models gold chalcopyrite
X(Bt*) tracking mix fluid dominance 1 0.9 0.8 0.7 Change in the mixing zone Background rock 0.6 X(Bt*) 0.5 0.4 0.3 0.2 0.1 0 Slice showing X Bt along the mixing zone 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Mass Fraction Brine
2D Grid simulations: Flow Through 2D grid of fluid flow through reactor time/distance relationships Distance increasing f/r is constant at any one time but time integrated f/r changes. Fluid source at constant flux Rock I x N 0 N 0 N x Time increasing Rock unit can contain multiple rock units so fluid can flow across geological boundaries Horizontal section Snap shot in time Vertical section Time evolution of a point in space N max Diagonal section evolution of infiltration front
Regional Albitisation Flow-through Infiltrate Na-modified granite fluid into Casilicates Fluid infiltration front is at step 21 Background dry rock beyond, saturated rock behind Volume% minerals 100% 80% 60% 40% 20% 0% Albite Cpx Korzhinskii Fronts Anorthite Act 1 2 3 4 5 6 7 8 9 10111213141516171819202122232425262728293031 Fluid infiltration to step 21 Distance along infiltration column Bt Ksp Qtz
Fluid chemistry in fronts Predicted pattern of Fe, Na and K in chemical fronts behind infiltration fornt What do FLINCS represent? Fe (molal) 5.0E-04 4.5E-04 4.0E-04 3.5E-04 3.0E-04 2.5E-04 2.0E-04 1.5E-04 1.0E-04 5.0E-05 0.0E+00 Total Fe(aq) Total K(aq) Total Na(aq) 0 5 10 15 20 25 30 35 Reaction Step 3.5E-01 3.0E-01 2.5E-01 2.0E-01 1.5E-01 1.0E-01 5.0E-02 0.0E+00 Na & K (molal)
469000E 470000E 471000E pmd CRC Alteration Association of regional Na-Ca alteration with ore-bodies Ore proximal K alteration 7739000N 7740000N Ernest Henry Diorite Intermediate volcanic rocks Calc-silicate, psammite, schist Minor (2-15%) Major (15-30%) Intense (> 30%) Potassic alteration (Kf ± mu, qtz, cc, py) Na ± Ca alteration 1 km projected position of albitite geochemistry traverse N 7738000N Rock boundary Fault or shear zone Breccia, cm- to dm-scale clasts Magnetite ± biotite alteration Ernest Henry orebody Ore related alteration
Log volumetric f/r pmd CRC Effluent models Effluent fluid run through pelites can produce alteration similar to proximal gangue assemblages in many of the Feox-Cu-Au deposits 100% 80% 60% 40% 20% 0% Mixing K-, Fe-enriched albitite effluent fluid with pelite after dropping some Ca as caclite veins Actinolite Magnetite Biotite Anorthite Albite Clinopyroxene Chlorite Muscovite Andalusite Rutile Biotite K-Feldspar Quartz 8.48 5.40 4.40 3.40 2.39 1.42 0.43-0.55-1.55-2.55-3.55
Linking GWB-HCh Developed software utilities that allow the same thermo dataset to be used for GWB and HCh. Plot high PT a-a and T-a diagrams to explore potential geochemical pathways (based on mineral assemblages) before going to fluidrock modelling. Linking software will become a web service towards the end of the year.
The Wallaby Example 1 2 3 300 o C 450 o C Po+(Py) Mt Whats the geochemical pathway? Mt-Py -(Hm)
Example from wallaby where GWB-HCh linking is proving an important work tool T ( C) 550 500 450 400 350 Pyrrhotite Magnetite Magnetite Hematite Hematite S-poor S-Rich? 300 Pyrrhotite Pyrite 250 Pyrite 200 10 5 0 5 10 log a SO 2 (aq)/h 2 S(aq) jc138398 Tue Jul 29 2003
Gold solubility (log m) pmd CRC 550-2 Magnetite 500-3 450 Pyrrhotite 400 350 300 H 2 S(aq) Hematite SO 2 (aq) T ( C) Diagram Fe ++, a [main] = 10 3, a [H 2 O] = 1, a [H 2 S(aq)] = 10 1 (speciates), ph = 4.5 (speciates), a [Au + ] = 10 5 ; Suppressed: Troilite -4-5 250 Pyrite 200-4.5 Sulphur 150 100-6 -7-5 -8-9 -10 10 5 0 5 10 log a SO 2 (aq)/h 2 S(aq) Jc140209 Tue Sep 02 2003
3D Gold Solubility Surface
Gold vectors to precipitation pmd CRC 550 Magnetite 500 450 400 350 300 T ( C) Diagram Fe ++, a [main] = 10 3, a [H 2 O] = 1, a [H 2 S(aq)] = 10 1 (speciates), ph = 4.5 (speciates), a [Au + ] = 10 5 ; Suppressed: Troilite Au-Cl 250 200 100 Pyrrhotite Au-OH Au-Cl complexes H 2 S(aq) Pyrite Sulphur 10 5 0 5 10 log a SO 2 (aq)/h 2 S(aq) Hematite Au-S complexes SO 2 (aq) Au-HS Au-HS We can now test different processes in reactor models which will account for all combinations 150 of vectors together Jc140209 Tue Sep 02 2003 0.001 m ΣS 3 m ΣCl
Future Challenges Software development that will increase the usability of geochemical modelling, including inputs, conceptualisation and data visualisation. Integrating fluid-flow (deformation) results as constraints on inputs/results. Integration of data to and from other pmd*crc projects to increase effectiveness of whole modelling process. More reality checking to better improve predictive nature of this type of modelling (where are we right, how mush have we predicted, where do we need to improve)?