Seeking advice on the tolerance settings for tearing flow strands in Aspen
Thread Content
There is a loop in the process, after which a tearing flow stream was created, and the convergence tolerance for this stream was customized. The calculation yields an error: * WARNING BLOCK R201 IS NOT IN MASS BALANCE: MASS INLET FLOW = 0.24064727E+02, MASS OUTLET FLOW = 0.23832831E+02 RELATIVE DIFFERENCE = 0.97301131E-02 THE IMBALANCE IS DUE TO A LOOSE TEAR TOLERANCE STREAM VAPOR TOLERANCE = 0.20000000E-01 This means that the tolerance setting for the VAPOR stream is too loose, while the calculation tolerance for module R201 is 0.009… I changed the convergence tolerance for VAPOR from 0.02 to 0.01; theoretically this should be sufficient to meet the tolerance requirements, but the calculation still results in an error. Due to mass imbalance, the data calculations are incorrect; I want to reduce this convergence tolerance but am not sure where to make the changes. {:3_65:} I hope an expert can give me some guidance…………
Different convergence problems have different solutions: either assign initial values to the tearing flow, or change the convergence method, and so on. It would be best to send over your BKP file for a look
Since the file relates to the project details, it’s really not possible to upload it. {:3_65:} You mentioned the issue of initial values; I tried it again today. First, I set a larger tolerance, then used the calculated values to replace the initial values for the tearing flow streams. After that, I reduced the tolerance and once again used the calculated values as initial values. The tolerance can be gradually reduced, but it still cannot be brought down to the desired value; if it is lowered, Err/Tol will reach around 5 and cannot be decreased any further. I see in Sun Lanyi’s textbook that if the value does not drop any further, the following measures can be taken: 1. Set a smaller tolerance for the internal iteration (the tolerance for BLOCK is 0.0001, while that for external tearing flow streams is 0.005; thus the external value is already higher than the internal one). 2. Increase the tolerance for the external iteration. 3. Use the Broyden or Newton method to achieve convergence for both the internal and external iterations. (I changed it to use Newton simultaneously, but an error occurred during calculation due to the lack of material properties = =) So I’m not sure how to adjust it; it seems that the only option is to use that tolerance. I’m unable to upload the file; the problem description might not be very clear. Still, thank you very much for your help. {:3_64:}
Unless you know exactly what you’re doing, Newon or Broyden are not recommended. Try increasing the number of iterations. Change the damping coefficient.
Can ASPEN simulate mass conservation? It’s basically the conservation of mass flow rate.
Okay, I’ll try changing the damping coefficient. I only know that this setting exists in the module; is there also a setting for the damping coefficient within the tearing flow section?