Impact of Simulation Methods on Column Relief Systems Analysis

Accurate required relief load calculation for column systems are critical in accurately sizing relief devices and relief load handling systems. Column systems, which can contain reboilers, condensers, pumparounds, refluxes, and other complexities, and have many overpressure scenarios that must be evaluated and for which many methods exist to estimate the required relief rate. The purpose of this article is to examine different relief estimation methods for a column system and discuss the results of each approach. These results are not meant to be prescriptive for future systems, but to provide additional information so that the most fitting approach can be used given the requirements and constraints of the project. Methods in this evaluation include unbalanced heat input, dynamic simulation, and several methods of steady state simulations of varying complexity from a single flash drum to a standard column simulation. The impact of rigorously modelling the column reboiler and overhead condensers vs using the normal operating duties was also evaluated. To simplify the analysis, I evaluate three common overpressure scenarios: (1) closed vapor outlets, (2) top tower reflux failure, (3) total power failure. Previous works in this area have compared the difference between the unbalanced heat input, steady state simulation, and dynamic simulation methods, but have not evaluated the impact of differing steady state simulation methodologies on the estimated relief rate. For the total power failure scenario relief rates were estimated by using both the normal reboiler duty and a rigorously modelled reboiler, as well as several different column liquids commonly used to evaluate this scenario. Simulations were performed using Aspen HYSYS V10 and existing process data for a Naphtha Stripper system. This article finds that for the evaluated system a series of flash drums closely matches the results of a standard column simulation. Decreasing the number of flash drums used resulted in a more conservative estimation. The results for the unbalanced heat input and dynamic simulation methods aligns with those of previous studies. The dynamic simulation method estimated lower required reliefs, while the unbalanced heat method is more conservative. Rigorously simulating the reboiler resulted in lower estimated relief for complex simulations, but when used with more simplified steady state methods was found to be more conservative than the unbalanced heat input method. Simulating the overhead condensers resulted in the largest variance for the closed outlets scenario, as at relief pressure the column overhead was fully condensed resulting in no relief. For the total power failure scenario, the dynamic simulation method resulted in the lowest estimated relief and was found to match closely with the results of the steady state method using the normal reboiler duty applied to the column feed liquid.