increasing number of steps

[AES]

Hello,

I am trying to predict the contribution that a covered stockpile will make to the pH of some nearby water in 30 years or so. The minerals of interest are in the react pane, but the amount of water flowing through the system is very small, thus the React program cannot converge on an answer.... how do I increase the number of steps or iterations?

Thank you!

[Brian Farrell]

Hi,

 

By default React will trace reaction paths in 100 steps, using a step size of 0.01 to advance the reaction progress from 0 (beginning) to 1 (completion). Under Config - Stepping..., you can change this default setting using the variable delxi. You might tell React to take smaller steps, say delxi = 0.001, so that the reaction path will be solved in 1000 steps.

 

You can change the maximum number of iterations taken at each step from the Config - Iterations... menu using the itmax (normal steps) and itmax0 (solving for initial system only) keywords. Section 2.7 (Settable variables) in the GWB Reaction Modeling Guide lists a number of additional variables that you might adjust for your runs. You can find more information about each of these variables in the GWB Reference Manual's section on React.

 

Hope this helps,

 

Brian Farrell

Aqueous Solutions LLC

[AES]

Hello Brian,

Thank you kindly for your reply. I played with delxi to get the program to converge on an answer (after 300000 steps!), but it didn't seem like the right answer- there were far too many minerals saturated, and the pH was very high. Thinking that we didn't take into account some clay phases that might be forming in the stockpile, I unsuppressed them in the next run. Now, once again, the program can't find a solution, and says the Xi step is too small. I am attaching the script because I wonder if there is something I am missing. Any insight you have would be much appreciated!

Thanks, Allison

 

Hi,

 

By default React will trace reaction paths in 100 steps, using a step size of 0.01 to advance the reaction progress from 0 (beginning) to 1 (completion). Under Config - Stepping..., you can change this default setting using the variable delxi. You might tell React to take smaller steps, say delxi = 0.001, so that the reaction path will be solved in 1000 steps.

 

You can change the maximum number of iterations taken at each step from the Config - Iterations... menu using the itmax (normal steps) and itmax0 (solving for initial system only) keywords. Section 2.7 (Settable variables) in the GWB Reaction Modeling Guide lists a number of additional variables that you might adjust for your runs. You can find more information about each of these variables in the GWB Reference Manual's section on React.

 

Hope this helps,

 

Brian Farrell

Aqueous Solutions LLC

stockpile_121912.txt

[Brian Farrell]

Hi Allison,

 

Increasing the number of steps is sometimes quite useful, but if you're getting up to 300,000 there may be some other issues which need to be addressed. If you attach the .rea file I can try to take a look at your problem.

 

Regards,

Brian

[AES]

Hi Brian

Thank you very much for your time.

Attached is the rea file.

Allison

Hi Allison,

 

Increasing the number of steps is sometimes quite useful, but if you're getting up to 300,000 there may be some other issues which need to be addressed. If you attach the .rea file I can try to take a look at your problem.

 

Regards,

Brian

stockpile_121912.rea

[Brian Farrell]

Hi Alison,

 

Do you know anything about the initial chemistry of your water (before reaction with the waste pile)?

 

Regards,

Brian

[AES]

Hi Brian, Thanks for your response. We set the initial water values at arbitrary very low concentrations for elements of interest so that we’d see them in the output file in “original basis”. The “true” initial water chemistry is the limited amount of rain water or groundwater which passes through a cover to interact with the mineralogy in the large stockpile. We are interested in determining the effect of water-rock interaction on subsequent pH and Ca, SO4, and Fe concentration in the seep water coming off the toe of the stockpile. We’ve characterized the major mineral contributions to porewater in the stockpile and are trying to use their reactivity and solubility to predict the end water chemistry.

I am wondering how realistic the output file is when I make delxi very very small- what is the effect of increasing number of steps on chemistry?

Thanks!

Allison

[AES]

Specifically, there is far more SO4 and Na and Mg that we expected to see in the output file... is that an artifact of having the reaction occur in so many steps?

Thanks!

Hi Brian, Thanks for your response. We set the initial water values at arbitrary very low concentrations for elements of interest so that we’d see them in the output file in “original basis”. The “true” initial water chemistry is the limited amount of rain water or groundwater which passes through a cover to interact with the mineralogy in the large stockpile. We are interested in determining the effect of water-rock interaction on subsequent pH and Ca, SO4, and Fe concentration in the seep water coming off the toe of the stockpile. We’ve characterized the major mineral contributions to porewater in the stockpile and are trying to use their reactivity and solubility to predict the end water chemistry.

I am wondering how realistic the output file is when I make delxi very very small- what is the effect of increasing number of steps on chemistry?

Thanks!

Allison

[Michal]

Hi Allison,

I am a newbie, but I think that the reason why the modeled concentrations are too high is due to the fixed O2(g) buffer that supply infinite amount of oxidant for pyrhotite oxidation.

 

I would consider using kinetic aproach and set up dissolution of pyrhotite and oxygen transfer from the atmosphere kineticaly. But the concentration will increase over time anyway. I think, using reaction transport in a column would be a good idea.

 

Kind Regards,

Michal

 

 

 

Specifically, there is far more SO4 and Na and Mg that we expected to see in the output file... is that an artifact of having the reaction occur in so many steps?

Thanks!

[Brian Farrell]

Hi Allison,

 

A few things. In general, if you have an analysis, including pH measurement, for the initial rainwater/ groundwater sample, I would use that as the starting point for my calculation (the initial system which you define in the Basis pane) rather than a completely arbitrary fluid composition. For example, if you're going to constrain pH by swapping the H+ component out for CO2(g) and setting its fugacity, you need a good idea of the total HCO3- component concentration (so that GWB can calculate the free HCO3- species activity, which fixes pH according the reaction H+ + HCO3- = CO2(g) + H2O). In this case, however, it looks like the huge amount of reactant minerals you add comes to dominate the chemistry eventually anyway. I'm not too sure what you mean by seeing the elements of interest in the output file.

 

As for changing the step size, ideally this shouldn't be too big of a factor in your results. Under normal conditions decreasing or increasing the step size can increase or decrease, respectively, the resolution of your model. The increasingly smaller steps taken by React here, however, indicate it is having a hard time converging on a solution. Try taking a look at the Mass H2O vs. Reaction progress. You'll find that all of your solvent water is used up, causing the run to crash.

 

As Michal states, as long as there is a large supply of O2(aq), Pyrrhotite will remain unstable and continue to oxidize, adding SO4-- to your system. This is not an artifact of the step size.

 

Hope this helps,

Brian