What is the impact of the Component Averaging Factor parameter in the PROII distillation column sure algorithm on the calculation results?
Thread Content
Dear experts, in the PROII distillation column sure algorithm, there is a parameter in the convergence data called Component Averaging Factor. What does this parameter mean? What impact does changing this parameter have on the calculated results? Will it cause inaccurate calculation results? I hope the experts here can help answer this. Thank you!
It is difficult for the tower to converge when using the default value of 0, but convergence occurs quickly when it is changed to 3. I wonder if changing this value has an impact on the calculation results?
This is a convergence aid under the SURE algorithm; it is somewhat similar to the damping factor in the IO algorithm, and it is not a requirement regarding convergence accuracy. Therefore, this setting may increase the number of iterations, but it has no impact on the convergence result. Below is the original information taken from the help files, provided for your reference: Component Averaging Factor: This weighting factor used for updating compositions is applicable to the Sure algorithm. A value of 1.0 assigns equal weight to the current and previous sets of compositions; a value of 2.0 gives twice as much weight to the previous set of compositions, and so on. PRO/II provides a default value of 0.0. Further Explanation: …… The component averaging factor for Sure columns is similar to the damping factor used for Inside-Out columns, and it can help bring columns with errors that fluctuate around a converged solution to convergence. The component averaging feature increases the number of iterations required for column convergence.
In my opinion, if the process converges, then the results should be fine. If possible, ASPEN can be used for comparison or to compare with actual field data. According to the guidelines, 3 should be the recommended upper limit for this value; since it hasn’t exceeded this limit, it’s still feasible. Component Averaging Factor (SURE algorithm only): A factor used to calculate the next set of trial compositions. A value of 1 assigns equal weight to the current and the previous set of compositions when they are averaged together. A value of 2.0 gives twice as much weight to the previous set of compositions. Values higher than 3.0 are not recommended.
Got it. Thank you all for your help; it’s been extremely useful.