(285b) A Valve Tray Design Capable of Efficient Operation in Tilted Columns and Out-of-Level Tray Columns

In this work, the operation and performance of a novel valve tray design capable of functioning without loss of efficiency when the column is operated in both vertical and tilted conditions is presented. The Tray Column used for the work is a Downcomerless liquid initiated and controlled Dual Flow Valve Tray with capacity to shut portions of the tray lacking liquid at any time. Portions of a Tray Column could be starved of liquid during operation when the tray is Out-of -Level and this may occur when Tray Columns are operated on moving platforms. The term “moving platform,” implies that the Column will not always be vertical, but can sway depending on the motion of the platform. These sways will result in tilting of the equipment mounted on the platforms, channeling of the fluids and a loss or total collapse of the efficiency of conventional Tray Columns subjected to such unique operating conditions. These conditions are encountered in many chemical processing operations that need to be carried out on floating platforms to recover “stranded gas” in remote locations, in stripping ships and even in columns mounted on land but exposed to strong wind forces and other adverse factors that cause the Tray column to be Out-of-Level.

A liquid controlled valve tray called “Plunger-Cap-Multifloat Valve Tray” having four floats per valve is examined. The “liquid control” aspect of the valve was achieved by exploiting the basics of the Archimedes principle of floatation and Upthrust. The fundamental aspect of this design was the determination of the optimum float size that will give the Upthrust required to operate the valve as desired. The design established that 0.5mm mild steel cylindrical floats with height/diameter of 50 mm will operate the valve when at least 75% of the float is submerged in water. The complete experimental rig used for this work consists of a 2-tray downcomerless column with 4 “Plunger-Cap-Multifloat Valves” per tray. Portions of the column shell were fitted with Perspex to permit observation of the workings of the tray internals during operation. Water supplied from a holding tank by a 0.5 Horsepower Electric pump was fed through the top of the column to the first tray. When the level of water on this tray was sufficient to lift the floats, the valves opened and allowed water to flow to the tray beneath. The response of the valves of the tray beneath was same as obtained with the tray above and the water flowed to the base of the column from where it was re-circulated by gravity to the holding tank. The air which was supplied by a 2.5 Horsepower Air Compressor was introduced beneath the second tray in the column. A head of liquid maintained at the base of the column served as a liquid seal and ensured that the air introduced flowed up through the open valve to the second tray first, then to the first tray contacting liquid in each of these trays. The air was released through a vent at the top of the column. The flow rates of both the water and air were measured by Rotameters. The column assembly was also fitted with valves for controlling the flows at the desired rates and the entire column is mounted on a rig which could be tilted and operated at various angles.

To evaluate the mass transfer performance of this column, an oxidation experiment was conducted and the percentage of Fe(II) oxidized to Fe(III) from the water by air at a fixed temperature  (30⁰C) was chosen as the system response. The impact of the various liquid and gas flow rates on this system response in both vertical and tilted conditions were examined.

Based on Factorial Design of Experiments, the effects of the Gas flowrate (x₁ in litres per minute), the Liquid flowrate(x₂ in litres per minute) and the Angle of tilt (x₃ in degrees) on the Response (y) were investigated. Both a First Order Linear experimental design and a Second Order Box-Wilson Rotatable design were used and four polynomial models tested.

The influence of these factors on the response was represented by the polynomial models and their effects were studied using the Student’s t-test and Fischer’s F-test for the Analysis of Variance (ANOVA) for each model. Out of the 4 models, the experimental data were found to be best represented by a linear polynomial model y = 12.33 + 1.30x₁ + 0.68x₂ + 0.235x₃ + 0.178x₁x₂ + 0.023 x₁x₃ + 0.043x₂x₃ with Coefficient of Determination R² = 0.9635. The test of significance of the individual coefficients in the model based on the Student’s t-test showed that only parameters x₁ and x₂ (i.e. the gas and liquid flowrates) are significant at 95% confidence level. The angle of tilt (with p=0.18) had no significant influence on the amount of Iron (II) oxidised. The ANOVA of this model showed that the model was significant as the calculated F-value (F-model = 17.60) exceeded the tabulated F-value (6.16), and also from its probability value (p=0.008) which is less than 0.05. From this test also, the model parameters x₁ and x₂ had significant influence as their calculated F-values (164.57 and 45.38 respectively) were much higher than their tabulated F-values (F=7.71). The parameter x₃ and their interaction effects were still not significant from this test as their calculated F-values (5.42 for x₃, 0.06 for x₁x₃ and 0.2 for x₂x₃) are less than their tabulated F-values of 7.71. These findings were further validated by the x-y scatter plots of the obtained data which showed that for all the liquid flowrates, the % Fe (II) oxidised increased with increasing gas flowrates but the impact of the angles of tilt was minimal. Also, for all the liquid flowrates within the range of our experiments, the change in the percentage Fe (II) oxidised between 0 degrees to 10 degrees column tilt and also between 0 degrees to 20 degrees column tilt is seen to be less than 1% at constant vapor rates.

The implication of these is that the operational efficiency of this novel Valve Tray Column will not be compromised by column tilt of up to 20 degrees from the vertical. This finding validates our earlier hypothesis that a liquid operated and controlled Valve Tray (such as the Plunger-Cap-Multifloat Valve Tray presented in this paper) will provide solution to the problems of liquid maldistribution and efficiency collapse occasioned by column tilt and motion.