Intoduction
An exothermic catalytic reactor is designed. Three reactions occur within this reactor of which two of the three reactions are side reactions. These side reactions is neglect in this reactor because the designed reaction is favoured at the operating temperature and pressure, only trace amounts of the by products are formed.
The amount of catalytic gauzes inside the reactor as well as the mass of the catalyst in determined. A kinetic model is chose to simulate the reactor in MATLAB. This model is used to determine the concentration profile of the components.
The reactor is equipped with cooling coils to control the reactor temperature. The physical parameters of the reactor, the diameter, length and wall thickness, is also determine.
Reactor Properties
The reactor has to accommodate a heterogeneous, catalytic reaction for a continuous process. The desired exothermic reaction is shown in Reaction1. This is a gas phase reaction and therefore it is assumed that the pressure does not change inside the reactor when the process is simulated in Aspen Plus®.
4NH3 + 5O2 ↔ 4NO + 6h3O (1)
The reactor is operated at a constant pressure of 4.9bar. The inlet temperature and outlet temperature is 150.5°C and 903°C, respectively. The energy released during the reaction is 905kJ and a conversion of 96.5% is reached. The platinum catalyst that is used in the reactor has a weight of 1.9kg per ton HNO3 produced in an hour. The mesh wire diameter of the platinum gauzes is 0.08mm. The activation energy is 125400kJ/kmol for this reaction (Sadykov et al, 1999).
The reactor feed consists of eight parts enriched air for every part ammonia. The enriched air contained 24% oxygen. The inlet molar flowrate to the reactor is 0.73kmol/s with an initial volumetric flowrate (Qo) of 8.22m3/ s.
The two side reactions occurring in the reactor is describing in Reaction2 and Reaction3. These side reactions together with their kinetic data are described in more detail in the group report (Fourie et al, 2009).
4NH3 + 3O2 ↔ 2N2 + 6h3O (2)
4NH3 + 4O2 ↔ 2N2O + 6h3O (3)
The Cp value of the process fluid is calculated with Equation4 with yA the mole fraction of the ammonia in the feed T is the temperature in K. This equation applies for the temperature region between 600K and 1100K and an ammonia mole fraction between 0.09 and 0.19. The mole fraction of ammonia in the feed stream to the designed reactor is 0.11 on a mole base (Rase, sa).
The exothermic heat of reaction is calculated with Equation5 in J/mol NH3 reacted across the same temperature region as the Cp value, while the viscosity () is calculated in poise, with T in K, according to Equation6 (Rase, sa).
The platinum catalyst used in the ammonia oxidation reaction consists of 90% platinum, 5% palladium and 5% rhodium (Rdzawski, Ciura & Nikiel, 1995). If a gas distributor is installed in the reactor it could reduce the mass of catalyst needed to approximately to 40% of the original catalyst mass (Rase, sa).
The reactor is sensitive to change in the ammonia volume % in the feed. If the ammonia feed exceeds the 14volume% mark an explosion will occur. The reactions occurring inside the reactor is sensitive to temperature and pressure changes. At low pressure and high temperature the desired reaction is favoured while all the side reactions are limited. For temperature above 423K - 473K N2 formation will be the favoured in the presence of the catalyst and N2O will start forming at 473K (Fourieet al, 2009).
Reactor Model
The reactions occurring in the reactor are very rapid; therefore it is assumed that there is no ammonia available at the catalyst surface. This means that the reaction cannot be limited with internal mass transfer (Rase, sa). Though, the reactor is controlled with external mass transfer at high temperatures; the reaction rate is limited by the rate of ammonia diffusion to the catalyst (Sadykov et al, 1999).
The kinetics of the ammonia oxidation on platinum gauzes has an effective first order rate constant of 24600s-1. The contacting time on the platinum gauzes is approximately 10-4s and the activation energy is 125.4kJ/ mol (Sadykov et al, 1999). This kinetic data is determent with a model describing the oxidation reaction of ammonia on a platinum gauze catalyst. Experimental work at 900°C and below 8atm with 10volume% ammonia in the feed stream has indicated that this is an accurate model.
The mixing model of the ammonia oxidation reaction is determent with the amount of CSTR's (Continuous Stirred Mixed Reactors) in series (n). The amount of CSTR's in series is determined to be 22CSTR's according to Equation8 and the help of Microsoft Office Excel. This equation depends on the mass of catalyst used in the reactor (W), the volumetric flowrate (Q) of the process fluid, and the rate constant (k') of the catalytic reaction. The Peckley number (Pe) for this reaction is calculated with Equation 9. The Peckley (Pe) number is 42; this implies that the mixing model can be presented with an ideal PFR (Plug Flow Reactor). The catalytic rate constant (k') is determent form the rate constant given in the literature by dividing with the catalyst density (Ccat), Equation10 (Nicol, 2009).
The PFR model that is used is shown in Equation11 and Equation12. In Equation 11 the mol balance of the ammonia component is shown. The kinetic rate expression is given by Equation12. This model is modified so that it enables one to determine the percentage conversion (x) of the ammonia on each catalyst gauze (G). It is determine that each catalyst gauze weighs about 0.57kg (this calculations is shown somewhere else in this report).This modified model is shown in Equation13 where Qo is the initial volumetric feed flowrate (Nicol, 2009).
A stoichiometric table is used to estimate the value of ε; this is shown in Table1. The numeric value of ε is 0.0175.
Stoichiometric Table
The amount of gauzes needed for a 96.5% conversion is determent in the next section of the report. The conversion of ammonia on the catalyst plates is shown in Figure1. According to the graph generated in MATLAB, only 36gauzes of catalyst are needed to achieve 96.5% conversion. This is confirmed with another graph generated in MATLAB where the conversion of ammonia is determent with the increase of catalyst mass.
According to both these figurers approximately two thirds of the catalyst is needed to achieve the desired conversion. According to Rase (sa) it has been shown that in high pressure plants only the upper two thirds of the catalyst bed are active; the lower third of the catalyst is not active. But when the extra (inactive) gauzes are removed the desired conversion will not be reached due to the increased pressure drop the reactor. This pressure drop is causes by uneven distribution of the process fluid as well as the bypass of certain portions of the catalyst (Rase, sa).
Reactor Design
The reactor diameter depends on the catalytic gauze structure. Typical reported catalytic requirements for reactions taking place at 900 °C and 100 psig is 80mesh catalyst gauze. The weight of 80mesh gauzes is 0.57kg/m2. The cross sectional area of these gauzes is 0.26m2 per 100ton HNO3 produced daily. The designed nitric acid plant produce 60% nitric acid at a rate of 25.25ton/h, this implies that 16.31ton/h pure nitric acid is produced. Therefore, the total weight of the catalyst needed is approximately 31kg and the area of the catalyst is 1m2 (a diameter of 1.13m). This implies that the minimum inside diameter for the reactor would be 1.13m. The 54catalytic gauzes are required, this is calculated with equation 14(Rase, sa).
The reaction does not depend on the volume of the reactor, therefore it is assumed that the reactor length will be equal to the cross sectional area of the reactor (1m). The volume of the reactor is 1m3 or 1000l. The length of the reactor should accommodate the catalytic gauzes, the platinum recovery gauze and the cooling coils. The internal layout of the reactor is described in more detail later in this report.
The reactor operates at 4.5bar while the atmospheric pressure is close to 1bar; therefore, the reactor is under internal pressure. The minimum wall thickness (t) is 7mm based on the circumferential stress. It is calculated with Equations15 (DeVaal, 2008).
The internal diameter (di) of the absorption column is 1.13m and the absorption column operating pressures (P) is at 4.5bar. The welding efficiency (Es) is of approximately 0.8 while the allowable stress (fallow) in the reactor is calculated with Equation16 with the corrosion allowance factor (c) of 3mm. The average diameter (dm) for the reactor is calculated with one of two equations, Equation17 or Equation18, depending on which diameter, the inside diameter (di) or the outside diameter (do), is available.
The internal diameter (di) of the reactor is 1.13m and therefore Equation18 is used to calculate the average diameter (dm) for the absorption column. It is found that the average diameter (dm) is 1.13m and allowable stress (fallow) is 79.8MPa.
Temperature Control
The temperature profile for an adiabatic reactor is shown in Figure3. The equation used to generate this profile is described in Equation19 (Nicol, 2009).
The reaction is an exothermic reaction, thus heat is released. The reactor is insulated to prevent heat loss due to conduction thru the reactor shell. The reason for the insulation is to protect the workers on the plant from the extremely hot outer surface of the reactor since the insulation will reduce the outer surface temperature of the reactor. But the reactor has to be cooled. If the reactor is not cooled in any way the temperature could rise as high as 5500K which would definitely melt the stainless steel shell and pipes of the reactor. The melting point of stainless steel is 1527°C (Perry & Green, 1997: 28-39).
Water could be fed to the reactor to absorb the reaction heat released during the execution of the exothermic reaction (Henckens & Scheibler, 1981). But then the water feed flowrate should be so small that it will only be present as a trace element in the feed composition. With Equation20 it is determent what the mass flowrate (m) of such a water stream should be. The Cp value of water is approximately 4.3kJ/kgK, the cooling water temperature is 20°C and the reactor have to be cooled down from 5227°C. It is assumed that the heat exchange occur with an efficiency of 85%, therefore the final temperature of the water will be approximately 784°C. The average Cp value of the process fluid, calculated with Equation4, is 13.48kJ/kgK and the amount of heat that has to be removed to decrease the process fluid temperature to 900°C is 1165.93MW. The water flowrate has to be 354.9kg/s. When this mass of water is added to the reactor feed it will have a significant effect on the process equilibrium and therefore the cooling water should rather flow through a pipe inside the reactor (Çengel, 2006: 135)
The heat transfer area (A) between the cooling water and process fluid needed is calculated with Equation 21. The overall heat transfer coefficient (U) for this process is 25kJ/kgK (Engineering page, 2009).
With the above specifications it is shown that the heat transfer area (A) is 10.69m2. The cooling water is sent through 3” stainless steel pipes with a total length of 2344m.
The insulation material has a density 240kg/ m3 and a thickness of 75mm and is known as mineral fibre insulation (Auralex, 2009). Mineral fibre is produced from slag wool which is constructed from melted industrial slag or steel mill slag that is spun into fibres. These fibres are treated with oil and binders to restrain dust particles and to maintain its shape. The fibres are then woven into buffs. This insulation material is chosen because it does not melt or support combustion. The thermal conductivity of the mineral fibre insulation is 0.04W/ m K (Çengel, 2006: 850).
Concentration Profile
The kinetic model as shown in Equation13 is used to determine the concentration profile of the components inside the reactor on each gauze. This concentration profile is developed with a MATLAB program and is shown in Figure4. The ammonia concentration is indicated with the dotted line while the oxygen concentration is indicated with the solid line with the same form as the dotted line. This is the concentration profiles of the reagents in the reactor. The concentration of the products formed in the reactor is illustrated with the other two lines. The dashed line indicate the concentration profile of the nitrogen monoxide that formed while the solid line with a similar form as the dashed line indicates the water concentration profile.
one can see that equilibrium is reached at approximately 36gauzes.
Pressure Drop
A broad study of the pressure drop across a wire gauze is presented by Armour and Cannon. The model is based on several experiments; the geometrical parameters of the wire gauze are taken into consideration during the development of the model. The Reynolds number (Re) range over which this model could be applied is from a Reynolds number (Re) of 2 to a Reynolds number (Re) of 700. The angle of the process fluid flow across the gauze (θ) is 21.3°while the porosity (ε) of the wire gauze is 0.67. The pressure drop across the laminar range is described by the Hagen-Poisseile equation and the pressure drop in the turbulent range is described by the Darcy- Weisbach equation. Therefore the modified Ergun equation is given in Equation22 (Kolodziej & Lojewska, 2008).
The pressure drop in Pa (∆P) across the reactor depends on the fluid density (ρ), the gauze wire diameter (dw) the superficial velocity (wo) in m/s given by Equation23 and the bed tortuosity factor () given in Equation24. The tortuosity factor is 1.19 for this specific platinum gauze.
The effective velocity (we) with a value of 8.2m/s is used to calculate a specific Reynolds number (Ree). This Reynolds number (Ree) with a value of 74.47 is used to determine the Fanning friction factor of the system form the graph shown in Figure5. From this graph the combined Fanning friction number (fapp + ft) is found to be 0.8 (Kolodziej & Lojewska, 2008).
With a reactor lent of 1m the total pressure drop across the reactor is 0.09MPa; this indicates a 20% pressure drop.
Internal Design
The ammonia oxidation reactor is not only equipped with the platinum catalyst gauzes and cooling coils, but also with platinum recovery gauzes. This filter like gauze can collect up to 60% of the catalyst losses (Brink, 2009). The catalytic catchment gauzes have to be cleaned and replace every now and again. To do this the top of the reactor have to be removed so that the platinum recovery gauzes can be replaces or removed for cleaning purposes.
When the reactor is started it is preheated to prevent a very rapid corrosive attack from the process fluid. The platinum recovery gauzes are located below the platinum catalytic gauze. Both these type of gauzes is situated below a perforated plate which will ensure a better gas distribution in the reactor. Tubes are installed just below the platinum catalyst gauzes; these tubes are the supersheater tubes which are used to preheat the reactor during start-up. Hydrogen burners in the burner head is often used to help with the preheating task.
The process fluid enters the reaction from the top, and then it is distributed by the perforated plate. The process fluid come in contacts with the catalyst where the actual reactions takes place before it flows through the recovery gauzes which catches most of the platinum particles that might have dissolved into the gas stream. The product stream leaves the reactor through a pipe at the bottom of the vessel. The cooling coils are installed at the wall of the reactor to cool it down and protect it from excessive temperature. These cooling coils are arranged spirally in disk form. An illustration of an example of the internal layout of the ammonia oxidation reactor is shown in Figure.6 (Thiemann, Scheibler & Wiegand, 2003: 18).
External Design
The reactor is 1 m in diameter and 1m in height. This reactor can safely be mounted on a concrete platform. This reactor is situated in a shed for the convenience of the workers, whether they have to replace the platinum catalyst, clean the platinum recovery gauzes or just do routine checkups.
This reactor is equipped with 1manway since the reactor is relative small. It is also equipped with a door that grant excess to the catalyst packing, this door is used to remove the unreactive catalyst from the top of the catalyst packing and replace it with new catalyst at the bottom of the catalyst packing; not all the catalyst is replace at ones since only the top two thirds are actively used as described before.
Conclusion
The reaction is controlled by external mass transfer and not at all by internal mass transfer.
The kinetics of the ammonia oxidation on platinum gauzes has an effective first order rate constant of 24600s-1. The contacting time on the platinum gauzes is approximately 10-4s and the activation energy is 125.4kJ/mol.
The mixing of this reaction is module with 22CSTR's in series and a Peckley number is 42; this indicates that an ideal PFR will be sufficient
The catalyst is a 80mesh gauzes with a density of 0.57kg/m2. Therefore, the total weight of the catalyst needed is approximately 31kg and the area of the catalyst is 1m2. The 54catalytic gauzes are required
It is assumed that the heat exchange occurs with an efficiency of 85%, and that a heat transfer area of 10.69m2. The cooling water is sent through 3” stainless steel pipes with a total length of 2344m.
The reactor is 1m in diameter and 1m in height and a wall thickness of 7mm.
The insulation material has a density 240kg/m3 and a thickness of 75mm and is known as mineral fibre insulation.
A 20% pressure drop across the reactor occurs.
This reactor can safely be mounted on a concrete platform. This reactor is situated in a shed
This reactor is equipped with 1manway since the reactor is relative small. It is also equipped with a door that grants excess to the catalyst packing.
