[cfx_ws1992]

My simulation is that the methane hydrate in porous media can decompose into water and gas,then water and gas can flow in porous media.After a series of hard work, it can run successfully, but I am not sure its convergence.So I have some questions.
(1)The pressure in the centre of the Porous,which is the calculation domain in the CFX model,has the tendency for increase after the Time Step of 30.My initial pressure is 3.75MPa,and the outlet pressure is 2.84MPa,the pressure should decrease with time.What causes this phenomenon?Is it due to the virtual wall at the outlet?
(2)Does the calculation of every time step require convergence?In other words,the calculation of every time step has completed before the Max.coefficient loop iteration reaches.In addition,when Max.coefficient loop iteration is reached, the computation converges forcibly.Whether it affects the outcome?
(3)The energy residual of the gas is about 8e-3,but the convergence criteria is1e-4.I don't know what caused it.
(4)How to evaluate convergence?The CFX HELP points out that the convergence is determined by the RMS of Monmentum and Mass,the RMS of Heat Transfer,Imbalance.The best outcome is that the RMS of Monmentum and Mass,the RMS of Heat Transfer,Imbalance of every phase (Gas,Water,Hydrate) are under the convergence criteria.But when One or two is over the convergence criteria,Whether it converges?And how we can adjust to change the misconvergence?For example,the U-Mom Imbalance, the RMS of Heat Transfer of Gas over the convergence criteria respectively,others under the convergence criteria respectively.What can we do?
Thank for helping !

[ghorrocks]

1. We cannot answer that without seeing the details of your model. But the forum is not suitable for detailed debugging like that, you are going to have to figure that out yourself.
2. If you want accurate time history then yes, you need to converge every time step.
3. Your simulation is having problems converging. You need to improve numerical stability by smaller time steps, double precision numerics or improve mesh quality. As this model involves porous materials you might also need to look at some options specifically for porous materials.
4. I suspect you are not using a suitable time step for your simulation. I recommend you change to adaptive time stepping, homing in on 3-5 coeff loops per iteration. Make sure the maximum and minimum time step size are wide enough that you never reach them. Then the simulation will find its own time step size and is more likely to converge. Then you will be able to run some simulations testing different convergence tolerances and work out what convergence tolerance you need in your case.

[cfx_ws1992]

1.I failed to upload the file yesterday which had a large memory.I am so sorry because I want to know some factors affecting the pressure calculation,such as the reference pressure,the settings of inlet and outlet .Anything else?
2.I just want a convergence result, regardless of the calculation process, so the calculation misconvergence of each time step has a little influence on the final result.May I think so?
3.I also know that the size of the time step is helpful to the convergence results.I change the time step of 0.001s,even 0.0001s .But the computing speed is another big problem.How to balance the two?
4.The difference between the two groups is Conservation Target and Residual Target.The hesternal one is Conservation Target = 0.01 and Residual Target = 0.005. The hodiernal one is Conservation Target = 0.5 and Residual Target = 0.05.

[ghorrocks]

1. What affects the pressure calculation? I think the only realistic answer I can give to that question is "everything".
2. I don't understand what you are asking. But the convergence shown in your plots is very bad and you are unlikely to have a well converged simulation. Please read this FAQ: https://www.cfd-online.com/Wiki/Ansy...gence_criteria
3. You need a time step small enough so you converge properly and have sufficient time resolution. If this results in too long simulation time then you need to get a more powerful computer.
4. I assume you are talking about the convergence plots you attached. None of those plots look converged, so none of your simulations are going to be usable so far.

[cfx_ws1992]

1. I do not care the calculation convergence,and I want to compare the temperature profile calculated by the computer with one observed by the experiment.
2.I try to use adaptive time stepping,but with the increasing of time step size,RMS shows an increasing trend and overs the setting RMS target.By comparing RMS profile and MAX profile,and refering to the imbalance graph ,I change time step of 1e-4s as the fixed time step.(RMS target=1e-4,MAX target=1e-3 )

[cfx_ws1992]

1.time step =0.0001s,Conservation Target = 0.5 and Residual Target = 0.0001
2.The cumulative time is a little short.

[ghorrocks]

I do not understand your comment "I do not care the calculation convergence,and I want to compare the temperature profile calculated by the computer with one observed by the experiment."

If your simulation is not properly converged then the results will not be accurate. So you need adequate convergence for an accurate simulation. That is why convergence is important.