NEW Thermal Case Study - Bespoke Modelling

This case study describes an analysis performed earlier this year on a Heavy-End re-boiler, which was identified with a potential thermal transient problem. The heat exchanger is normally used for heating a flow of hazardous liquor with a pressurised steam flow. The vessel is a horizontal cylinder with 2 dished ends which create end plenums, capped by 2 tube-sheets 9m apart. These mount a bundle of straight longitudinal tubing, passing the liquor along the top half and back along the bottom half. Normally, steam flows in between the two tube-sheets from the hazardous liquor inlet end past a 'slalom' pattern of baffle plates with missing segments on alternating sides. Condensation outlet and vents are at the right hand end. One mode of operation is for the steam flow to be replaced by water at nominally ambient pressure and temperature. Concern existed about the effects of thermal shock due to the sudden cooling of the pipes. The objective of the analysis was to quantify the safe maximum water flow rate that can be suddenly introduced without failure, in particular with regard to maintaining the integrity of the tubing end connections. Three main problems were envisaged: 1. Sudden cooling of the pipes occurs directly beneath the water inlet but over only the length between the adjacent tube-sheet and the first baffle plate. 2. The hazardous liquor underneath the top tubes is cooled by the water flow and carries this 'coolth' down the pipes. The water will flow mostly down and along the bottom half of the vessel, cooling some of the lower pipes by diminishing amounts. 3. The case of cooling water being admitted when the piping is empty may also be possible, so is subjected to the same method as above. Several pieces of physics needed to be accounted for: - Mass advection of the hazardous liquor, water and steam down and over the pipes. - Heat conduction through and along the pipes. - Convective heat transfer inside and outside the pipes. - Changing thermal and mechanical mass of the internal fluids. - Thermal expansion of the top and bottom piping bundles and the vessel. - A complicated statically indeterminate stress state with the vessel contracting and rotating at the bellows and sliding along a pad. This is beyond the ability of most FE codes to model, requiring thermal and mechanical FEA coupled basic CFD. Instead, a 1D scheme was built in MathCAD, accounting for liquid on either side of the pipe. Transient cooling curves were computed showing how the cooling water heated up and boiled whilst the hazardous liquor cooled. The thermal movement of the vessel was calculated along with the force histories on the pipes, showing that in fact the equilibrium case was critical. The allowable flow rate for normal hazardous liquor flow was shown to be very high, and this removed the process concern about the upper-bound allowable water flow rate. However, if the hazardous liquor flow is zero and the pipes are empty then this was shown to be unacceptable for any water flow rate if that flow is sustained. This is because the hazardous liquor flow keeps the bottom pipes hot, reducing the equilibrium stress state. The pipe end weld stress increases until failure after 8.2 seconds, so very short durations of water inflow may be admissible, or else allowing the vessel will fill up with water to well over half full in less than 8.2 seconds (bathing the top half of the pipes in relatively cool water). This however requires a greater understanding of the vessel?s overall discharge rates in the context of the process, and this was suggested as a subsequent activity. This study demonstrates the need for a bespoke approach to some problems, and how this can provide insight that off-the-shelf analysis cannot.

For more information on NEW Thermal Case Study - Bespoke Modelling talk to DC White Engineering Consultants