Unsteady State Operation in Trickle Bed Reactors “Modulation of input variables or parameters to create unsteady state conditions to achieve performance better than that attainable with steady state operation” Motivation and Objectives Unsteady State Operation in Trickle Bed Reactors “Modulation of input variables or parameters to create unsteady state conditions to achieve performance better than that attainable with steady state operation” Motivation and Objectives n Performance enhancement in existing reactors n Design and operation of new reactors n Lack of systematic experimental or rigorous modeling studies in lab reactors necessary for industrial application n Experimentally investigate unsteady state flow modulation (periodic operation) for a test hydrogenation system n Develop model equations incorporating multiphase, multicomponent transport that can simulate unsteady state operation CREL

Strategies for Unsteady State Operation n Flow Modulation (Gupta, 1985; Haure, 1990; Lee and Silveston, 1995) –Liquid or gas flow –Liquid/gas ON-OFF or HIGH-LOW flow –Isothermal/non-isothermal/adiabatic conditions n Composition Modulation (Lange, 1993) –Periodic switching between pure or diluted liquid/gas –Quenching by inert or product (adiabatic) n Activity Modulation (Chanchlani, 1994; Haure, 1994) –Enhance activity due to pulsed component –Removal of product from catalyst site –Catalyst regeneration due to pulse CREL

Gas Limited Reactions n Partial Wetting of Catalyst Pellets -Desirable –Internal wetting of catalyst –Externally dry pellets for direct access of gas –Replenishment of reactant and periodic product removal –Catalyst reactivation Liquid Limited Reactions Liquid Limited Reactions n Partial Wetting of Catalyst Pellets-Undesirable –Achievement of complete catalyst wetting –Controlled temperature rise and hotspot removal Possible Advantages of Unsteady State Operation CREL

Test Reaction and Operating Conditions Operating Conditions Superficial Liquid Mass Velocity : kg/m 2 s Superficial Gas Mass Velocity : 3.3x x10 -3 kg/m 2 s Operating Pressure : psig (3-15 atm) Feed Concentration : % ( mol/m 3 ) Feed Temperature : o C Cycle time,  (Total Period) : s Cycle split,  (ON Flow Fraction) : Max. Allowed Temperature Rise : 25 o C Alpha-methylstyrene hydrogenation to isopropyl benzene (cumene) CREL , (sec) s (1-  )  L(peak) L (base) L(mean)

Liquid Limited Conditions (0.4 <  ) High Pressure, Low Liquid Feed Concentration Gas Limited Conditions (  ~ 20) Low Pressure, High Liquid Feed Concentration CREL Comparison of Performance under Gas and Liquid Limited Conditions

Effect of Cycle Split and Total Cycle Period on Performance Enhancement Gas Limited Conditions (  ~ 20) Operating Conditions : Pressure=30 psig Cycle Split (  )= Liquid ON Period/Total Cycle Period(  ) CREL

Effect of Liquid Mass Velocity and Total Cycle Period on Unsteady State Performance Enlargement of enhancement zone at lower mass velocity , (sec) s (1-  ) L (ON) s (1-  ) L (ON)=L(mean) /  CREL L (mean)

Effect of Liquid Reactant Concentration on Performance Enhancement Lower conversion at higher feed concentration reduces enhancement even at lower liquid mass velocity , (sec) s (1-  ) L (ON) s (1-  ) L (ON)=L(mean) /  CREL L (mean)

Effect of Pressure on Steady and Unsteady State Performance At low mean liquid mass velocity, unsteady state performance is higher than steady state even as liquid limitation is reached  ~ 24  ~ 3 CREL , (sec) s (1-  ) L (ON) s (1-  ) L (ON)=L(mean) /  L (mean)

Effect of Cycling Frequency on Performance Optimum cycling frequency depends upon feed concentration, pressure and cycle split CREL

Effect of Base-Peak Flow Modulation on Performance Enhancement under Liquid Limited Conditions , (sec) s (1-  )  L(peak) L (base) At high peak to base flow ratio, unsteady state operation gives better performance even under (near) liquid limited conditions (0.4<  < 2) CREL L(mean) =  *L (peak)+(1-  )*L (base) L (mean)

Phenomena occurring under unsteady state operation with flow modulation in a trickle-bed reactor GOAL : To Predict Velocity, Holdup, Concentration and Temperature Profiles CREL

The Model Structure z=L z=0 GAS LIQUID SOLID C 1G C 2G. C nG C 1L C 2L. C nL Ni GS Ni LS Ni GL E GS E LS E GL Ni GL Bulk Phase Equations Species Energy CREL

Advantages of Maxwell-Stefan Multi-component Transport Equations over Conventional Models Advantages of Maxwell-Stefan Multi-component Transport Equations over Conventional Models n n Multicomponent effects are considered for individual component transport [k]’s are matrices n n Bulk transport across the interface is considered N t coupled to energy balance (non zero) n n Transport coefficients are corrected for high fluxes [k] corrected to [k o ] = [k][  [exp([  ])-[I]] -1 n n Concentration effects and individual pair binary mass transfer coefficients considered n n Thermodynamic non-idealities are considered by activity correction of transport coefficients n n Holdups and velocities are affected by interphase mass transport and corrected while solving continuity and momentum equations CREL

Flow Model Equations u iL, u iG  L,  G,P Staggered 1-D Grid Z Momentum Continuity Pressure CREL

Stefan-Maxwell Flux Equations for Interphase Mass and Energy Transport Gas-Liquid Fluxes Liquid-Solid and Gas-Solid Fluxes Bootstrap Condition for Multicomponent Transport Interphase Energy Flux for the Gas-Liquid Transport and Bulk to Catalyst Interphase Energy Flux for the Gas-Liquid Transport and Bulk to Catalyst Interface Transport Interface Transport Net Zero Volumetric Flux for Liquid-Solid and Gas-Liquid Interface for Net Zero Volumetric Flux for Liquid-Solid and Gas-Liquid Interface for Intracatalyst Flux Intracatalyst Flux CREL

Catalyst Level Equations Approach I: Rigorous Single Pellet Solution of Intrapellet Profiles along with Liquid-Solid and Gas-Solid Equations Approach II: Apparent Rate Multipellet Model Solution of Liquid-Solid and Gas-Solid Equations G C iCP L xc C iCP L G L G L G Type I: Both Sides Externally Wetted Type II: Half Wetted Type III: Both Sides Externally Dry CREL

Liquid Holdup and Velocity Profiles Operating Conditions: Liquid ON time= 15 s, OFF time=65 s Liquid ON Mass Velocity : 1.4 kg/m 2 s Liquid OFF Mass Velocity: kg/m 2 s Gas Mass Velocity : kg/m 2 s CREL

Transient Simulation Results Alpha-methylstyrene Concentration Profiles Alpha-methylstyrene Concentration during ON cycle of flow modulation Feed Concentration : 1484 mol/m 3 Pressure : 1 atm. Reaction Conditions : Gas Limited (  ~ 25) (Intrinsic Rate Zero order w.r.t. Alpha-MS) CREL

Transient Cumene and Hydrogen Concentration Profiles CREL Profiles show build up of cumene and hydrogen concentration during the liquid ON part of the cycle

Alpha-methylstyrene and Cumene Concentration Profiles During Flow Modulation Supply and Consumption of AMS and Corresponding Rise in Cumene Concentration Operating Conditions: Cycle period=40 sec, Split=0.5 (Liquid ON=20 s) Liquid ON Mass Velocity : 1.01 kg/m 2 s Liquid OFF Mass Velocity : 0.05 kg/m 2 s Gas Mass Velocity : kg/m 2 s CREL

Catalyst Level Hydrogen and Alpha-methylstyrene Concentration Profiles During Flow Modulation Concentration of Hydrogen during Liquid ON (1:20s, Wetted Catalyst ) and Liquid OFF(20:40 s, Dry catalyst) for negligible reaction test case Concentration of Alpha-MS in previously dry pellets during Liquid ON (1:20s, Wetted Catalyst ) and Liquid OFF(20:40 s, Dry catalyst ) CREL

Simulated Cycle Time and Cycle Split Effects on Unsteady State Performance Cycle Split and Cycle Period Effects Agree Qualitatively with Experimental Results CREL

Transient Fluid Dynamic Simulation using CFDLIB (Los Alamos) 2D-Test bed Dimensions: 29.7x7.2 cm 33x8 (264 cells with preset porosity) Cycle period= 60 s Cycle split = 0.25 Liquid Velocity = 0.1 cm/s (central point source) Gas Velocity= 10 cm/s (uniform feed) Gas-Solid and Liquid-Solid Drag Closure: Two-Phase Ergun Equation CREL Bed Porosity (lighter areas: higher porosity)

Liquid Holdup Comparison between Steady and Unsteady Operation t= 15 s t= 25 s t=40 s Steady State Unsteady State CREL

Summary Summary Performance enhancement was seen to be a strong function of the extent of reactant limitation Performance enhancement under gas limited conditions was found to be significantly dependent upon the cycle split, cycle period, liquid mass velocity and cycling frequency Performance enhancement under liquid limited conditions was observed only with BASE-PEAK flow modulation (to a lesser extent than under gas limited conditions) Rigorous modeling of mass and energy transport by Stefan-Maxwell equations and solution of momentum equations needed to simulate unsteady state flow, transport and reaction has been accomplished. Qualitative comparison with the experimental observations has been successfully demonstrated. The developed code can be used as a generalized simulator for any multicomponent, multi-reaction system and can be converted to a multidimensional code for large scale industrial reactors Fluid dynamic codes (CFDLIB) have been used to demonstrate better flow distribution under unsteady state operation. These codes would help achieve quantitative predictions when used in conjunction with the reaction transport simulator developed in this study CREL

Recommendations for Future Work n Downflow and Upflow Comparison Generalization of the conclusions obtained for complex reactions n Steady and Unsteady State Models Implementation for multi-reaction problems and conversion to a flow sheet based package (ASPEN user model) sheet based package (ASPEN user model) Implementation for multi-dimensional test cases in the framework of CFD codes (CFDLIB or FLUENT) CFD codes (CFDLIB or FLUENT) n Unsteady State Experiments Testing of reaction networks for possible enhancement in selectivity via flow or composition modulation flow or composition modulation CREL

Advisors: Prof. M. P. Dudukovic and Prof. M. Al-Dahhan Committee members: Prof. B. Joseph, Prof. R. A. Gardner Dr. M. Colakyan (Union Carbide) Dr. M. Colakyan (Union Carbide) Dr. R. Gupta (Exxon Research) Dr. R. Gupta (Exxon Research) CREL Industrial Sponsors Dr. Kahney, Dr. Chou, G. Ahmed (Monsanto) Dr. Patrick Mills (Du Pont) Engelhard, Eastmann Chemicals CREL Students and Research Associates Y. Wu, Y. Jiang Computer and Laboratory Support Dr. Y. Yamashita, Dr. S. Kumar, S. Picker, J. Krietler Parents, Roommates, and Friends Acknowledgements Acknowledgements CREL