using alphaMELTS (PTpath mode)

[sizovae]

Dear Paula,

May I ask you about a possibility to use alphamelts (MELTS and pMELTS) for the following:
I would like to get the stable mineralogy for such lithologies like basalt and peridotite, resolved on a grid of pressure and temperature (the temperature diapason~: 573 K - 1773 K; pressure: 0,5 kbar - 50 kbar). At the end I need to extract phase relations (including melt), composition of the melt, density, and enthalpy for each point of this grid.
Is that possible to do this using alphaMELTS?

Thank you very much.

Best regards,
Elena

[Paula]

Dear Elena,

Sorry for the delay in replying. You can certainly calculate the stable mineralogy on a grid using alphaMELTS. Text output includes: mass and composition of all solid and liquid phases; thermodynamic information, such as enthalphy and volume, for each phase; density of the melt; density of the residue (and you can calculate it for individual phases from the information given).

However, the calibrated range of the MELTS and pMELTS models do not extend as far as what you are aiming for. The MELTS liquid model is calibrated on a wide variety of natural bulk compositions and is suitable at pressures not greater than 30 kbar (ideally <= 20 kbar), and temperatures >= 773 K. The pMELTS liquid model is intended for peridotite and pyroxenite bulk compositions only, between 10 and 40 kbar (ideally <= 30kbar) and for temperatures >= 1273 K. At the lower subsolidus temperature end there tends to be problems with phases that have ordering effects (e.g. spinel, opx), which can cause the alphaMELTS program to hang.

One of the easiest way to do calculations on a grid is to use a PTfile; see ALPHAMELTS_PTPATH_FILE in the documentation. You will have the most success if you arrange the P-T pairs so the calculation runs down temperature at each pressure. Change the pressure by a fairly small increment, jump the temperature up to the maximum value, and repeat. Also check out entries for ALPHAMELTS_DELTAP, ALPHAMELTS_DELTAP, and ALPHAMELTS_SKIP_FAILURE.

Best,
Paula

[sizovae]

Dear Paula,

Thank you very much for the reply. I'll try.

Best regards,
Elena Sizova

[sizovae]

Dear Paula,

Could you be so kind to help me with few errors at running alphamelts?

I am trying to make calculations using PTpath. I have created a ptbeg.txt file (T=from 1000 up to 1700; P=from 10000 up to 10400; just to try) and started the run_alphamelts. But an error in MELILITE.C (function ORDER) occurred at some PT conditions, so the program, so Failed to converge!. Please, see an attached file (error in melilite).

Could you, please, tell me what kind of error is that?

Another question is where I am trying to produce the calculations at low temperature and pressure (T from 500 and P from 500 bar) using MELTS, after "Norm calculation of initial guess" the alphamelts is crashed, one can not do anything anymore; needs to restart the program.

Thank you in advance!

Best regards,
Elena   

[asimow]

Hi Elena,

I'll be handling most of the forum stuff for Paula for a while.

The MELILITE error is not a big problem. We don't actually care about melilite for this calculation. So, two work-arounds: (1) in your input file, suppress this phase (Suppress: melilite); and (2) in your environment variables file, set the option ALPHAMELTS_SKIP_FAILURE so that the code will continue past errors like this.

The crash at very low temperature can arise from several problems; 500 °C is lower than the program was ever intended to be used. Recommend for a PTgrid calculation that you start with the highest temperature first, and a superliquidus initial guess. Subsequent grid points will inherit the initial guess from the previous solution. It should have no problem with the upper temperature part of the grid, and as it comes down temperature at some point it will fail -- but hopefully at a lower temperature than you actually need results!

-- Paul

[sizovae]

Thank you  very much! It works.

[sizovae]

Dear Paul,

I have got a problem with a calculation of PT path. The minimum T = 1000 C; P = 10000 bar; maximum T = 1673.92444 C; P = 39871 bar; step for T = 3.43839 K; for P = 199.14 bar. During calculations for some P-T conditions numerous lines with "Bound check..." appear and at the end an error is THE QUADRATIC MINIMIZATION ALGOROTHMAS FAILED O CONVERGE, failure in silmin. The calculation continues further since I have ALPHAMELTS_SKIP_FAILURE true in the setting file. But I can not get results for these conditions with errors (these points are missing in the output file (Phase_mass_tbl.txt)). A strange thing for me is that in case I repeat these missing points separately they are normally calculated. Could you, please, write me, what might be a problem for that?

Thank you in advance.
Best regards,
Elena

[asimow]

Hi Elena,

I'm sorry, I just noticed your message has been sitting around un-answered for months. Do you still need help with this PTPath calculation? If so, please attach all three input files: your composition input file, your environment variable settings file, and your PTPath file. Then I can try to reproduce your error and see how to work around it...

-- P