CHAPTER 4 EXPERIMENTAL METHODS

Transcription

1 47 CHAPTER 4 EXPERIMENTAL METHODS 4.1 INTRODUCTION The experimental procedures employed in the present work are oriented towards the evaluation of residence time distribution in a static mixer and the estimation of kinetic parameters for the major hydrotreating reactions. The experimental data is also used to validate the three phase heterogeneous model developed in the present work and simulate the performance of static mixer over a wide range of operating conditions. Further, the kinetic data obtained from static mixer experiments is used to apply the model to industrial reactor. The details of static mixer experiments and the analysis of feed and product samples are discussed in this chapter. The object of this experiment is to study the behaviour of a plugflow reactor by performing a series of experiments on the saponification of Silver Nitrate Solution. 4.2 TUBULAR REACTOR In a tubular reactor, the feed enters at one end of a cylindrical tube and the product stream leaves at the other end. The long tube and the lack of provision for stirring prevent complete mixing of the fluid in the tube. Hence the properties of the flowing stream will vary from one point to another, namely in both radial and axial directions.

2 48 In the ideal tubular reactor, which is called the plug flow reactor, specific assumptions are made about the extent of mixing: 1. No mixing in the axial direction, i.e., the direction of flow 2. Complete mixing in the radial direction 3. A uniform velocity profile across the radius. The absence of longitudinal mixing is the special characteristic of this type of reactor. In an ideal PFR, a pulse of tracer injected at the inlet would not undergo any dispersion as it passed through the reactor and would appear as a pulse at the outlet. The degree of dispersion that occurs in a real reactor can be assessed by following the concentration of tracer vs time at the exit. This procedure is called the stimulus-response technique. The nature of the tracer peak gives an indication of the non-ideality that would be characteristic of the reactor under study. Thus the objective is to observe the dispersion that occurs in a Tubular Reactor by means of a tracer study and to determine the Time Distribution by the Stimulus-Response technique. 4.3 EQUIPMENT DESCRIPTION Tubular and packed bed reactors are commonly used for continuous operation to simulate plug flow conditions. However, for large systems, considerable axial and radial dispersion exist, producing deviations, from plug flow conditions. Use of internals to control the residence time distribution and or to improve heat addition/removal is common. The present investigation incorporates the use of a specially designed internals to achieve plug flow conditions. The equipment consists of an outer pyrex tube of 1210 mm in length and 35 mm inner diameter provided with two inlets at one end opposite to each other and inclined at 45 with the axis of the tube. The inlets are 15 mm in diameter. Four intermediate outlets are provided 212 mm apart, along

3 49 the length of the tube. These exits serve as manometer tappings for pressure drop experiments and would be used to turn out reactor effluent at intermediate length of the tube when conversion experiments are carried out. Alternatively, they would also be used to insert thermometer with slight modification, when heat transfer experiments are conducted. Those exits not in use are normally kept closed by spring-type pinch cocks on flexible rubber tube fitted on to the side exits. Figure 4.1 Tubular Flow Reactor with Helical Mixing Elements

4 50 Concentric with the outer pyrex tube is an inner tube 22.4 mm outer diameter 18.5 mm inner diameter. A brass flange brazed to the tube is bolted to a corresponding perspex flange, fastened to the pyrex tube with an adhesive. Figure 4.1 shows the unit with mixing elements. To the brass tube the helical elements are brazed. Each element consist of two 0.8 mm brass sheets 6.2 mm wide, carefully cut and twisted to form part of a helix 60 mm pitch. The elements are half a pitch in length i.e. 30 mm. These are brazed to the outer circumference of the tube 180 apart, alternative elements are righthanded and left-handed and each one is at right angles to the adjacent ones. In all, the reactor contains thirty five elements seventeen left-handed and eighteen right-handed. Figure 4.2 shows the experimental setup of plug flow reactor. Figure 4.2 Plug Flow Reactor The action of the mixing elements is somewhat similar to that of the static inline-mixer. The streams entering the reactor get split into two at the first element. Each split streams encountering the helix is twisted clockwise (If the element is right handed) and at the next element it is twisted anti-clockwise (if the element is left-handed).

5 51 The stream splitting and recombination contributes to effective mixing between fluid elements, as they travel along the reactor. The split layers are also constantly moved back and forth with respect to each other since the helical path length travelled by the fluid element depends upon the radial distance from the centre. This is referred to as layer-layer displacement. Apart from this, there is mixing action caused due to the swirling motion induced by the helices. An additional feature of the reactor is that the mixing elements would also serve as fins to improve efficiency of heat transfer when heat transfer medium flows through the annular core. An identical annulus, without helical elements is also constructed to compare the performance of the static mixer. The new type of this reactor with mixing elements is that on which experiments such as residence time distribution, is conducted. The experiments revealed that the annulus with mixing elements is a good approximation to a plug flow reactor even at low flow rates, since the helical elements act as turbulence creators. The equipment being similar in construction to the static in-line mixers can be used for efficient solid-solid mixing and for extrusion. The equipment can be put to use for any physical mixing process, where a high degree of mixing efficiency is required. The residence time distribution (RTD) for the unit is determined by measuring the response to a near pulse input. Water from the over head tank is allowed to flow through the unit. When the steady state is reached, the tracer is introduced into the reactor inlet. A hypodermic needle No.22 is used to inject 1 ml of tracer (nearly 5N Sodium Chloride solution) in the reactor inlet. Samples of reactor effluent are collected at frequent intervals (of 5 seconds). From the quantity of sample collected, a small quantity of sample (10ml) is titrated against a 0.05 N silver nitrate solution. The volume of silver nitrate consumed is directly proportional to the concentration of sodium

6 52 chloride solution in the outlet stream. The sample collection is done for sufficient time interval (for t>5e, where E is the average retention time). The above experiments are repeated until concordant titre values are obtained. The flow rate is then varied and the above mentioned procedure is repeated. For the annulus with mixing elements the flow rate range covered is litre/hr, as above 120 litre/hr, the average retention time was very low to detect any appreciable change in concentration within the time intervals of collection period. The above experiments are conducted in a similar manner in the annulus without mixing elements. Hence, it is observed that the variance is low for annulus with mixing elements than the simple annulus and further the range is from 0.12 to EXPERIMENTAL PROCEDURE The reaction chosen is the saponification of Silver nitrate in aqueous solution, with sodium chloride at room temperature. The reaction has shown to be bimolecular and second order according to the following stoichiometric equation, NaCl + AgNO 3 3AgCl + NaNO 3 (4.1) For each series of runs, the operational procedures are as follows 1. Prepare 7 litres of 0.05 N Silver nitrate solution and 5N Sodium chloride solutions. Then pour these solutions into the feed tanks. 2. Adjust the temperature of PFR. 3. Adjust the apparatus to constant flow rate. 4. Record the conductivity when steady state is reached. 5. Repeat the same procedure for different flow rates.

7 53 An identical annulus, without helical elements is also constructed to compare the performance of the static mixer. 4.5 DISTRIBUTION OF RESIDENCE TIME Using the distribution of residence time distributions for the analysis of chemical reactors was first proposed as far back as 1935 (Fogler 1999). However the concept of residence time distributions were further developed by Danckwerts (1953) when he gave structure by defining the various distributions of interest. The RTD of a reactor is a characteristic way of identifying the mixing occurring within a chemical reactor. For example, an ideal plug flow reactor exhibits no axial mixing and this is illustrated in the shape of the residence time distribution. Not all residence time distributions are unique to a certain type of reactor (Fogler 1999) and therefore different reactors can yield the same residence time distribution Measurement of the RTD The RTD is determined experimentally by injecting a chemical that is inert called a tracer, into the reactor at some initial time. The concentration of the tracer is then measured in the effluent stream as a function of time. To ensure that the tracer truly reflects the behaviour of the reactant mixture flowing through the reactor, it is important that the tracer have certain characteristics such as: Having properties similar to the reactor mixture Be soluble in the reacting mixture Should not absorb on the surfaces of the reactor.

8 54 The method of injecting the tracer is very important for the determination of the RTD. The two most used methods of injection is a pulse input and a step input. Injection with a pulse input is briefly described, as this is the method that will be used in the experimental work. With a pulse input an amount of tracer is injected into the feed stream entering the reactor in a very short period of time. be defined. For a pulse input the residence time distribution function, E( can QC( E ( (4.2) N 0 In equation (4.1), Q is the volumetric flow rate of the effluent, N 0 is the amount of tracer that was injected into the reactor and C( is the effluent concentration-time curve, known as the C curve in RTD analysis. The function E(, also known as the exit-age distribution function, describes the time spent in the reactor by different fluid elements. Usually N 0 is not known directly and can be obtained by the summation of all tracer amounts between the times the tracer was injected up to infinity. In addition, the volumetric flow rate is usually constant. From (4.2), the RTD function can therefore be defined differently, as in (4.3). E ( C( (4.3) C( dt 0 concentration curve. The integral in the denominator of (4.3) is the area under the It is important to state that drawbacks do exist when the method of injection is by means of a pulse input. Difficulties of the pulse input lie with

9 55 obtaining a reasonable pulse at the entrance of the reactor. It must be ensured that the time or period of injection be small when compared to the residence times of the chemical reactor. In addition, there must be, at the most, a small amount of dispersion occurring between the point of injection and the entrance to the reactor, if these aspects can be overcome, the RTD can be obtained simply and directly. In the case of certain types of reactors, especially in static mixer, a long tail is observed in the concentration-time curve. This occurrence can create large inaccuracies in the analysis of the RTD function. The tail in the concentration-time curve affects the denominator in 4.3, i.e. the integration of the C( curve. To overcome this, it has been suggested by authors to extrapolate the tail and analytically continue the calculation or the tail can be approximated as an exponential decay (Fogler 1999) Important Features in RTD Many features in RTD have been described in various articles (Naor and Shinnar 1963; Robinson and Tester 1986) and are: Measure of Spread The spread in residence times is an important feature of interest as it gives an indication of the deviation from plug flow in the flow system. An indication of the spread in residence times is done through the coefficient of variation or otherwise known as the relative standard deviation (Naor and Shinnar 1963). The standard deviation is defined as V t t E E( dt 2 1 (4.4) 0 where t E is the apparent residence time of the outlet concentration profile.

10 56 by: F( is the external cumulative residence time distribution defined t F( E( dt (4.5) 0 Internal Time Distribution. It is useful to use distribution functions that describe the internal conditions of a system. Such a distribution is known as the internal residence time distribution function. This function is related to the interior flow pattern and was first defined by Buffham (1983) as: t ( t ) E( (4.6) The internal RTD is simply also determined from inlet-outlet tracer experiments and does not provide additional information above the external RTD, E(. The internal RTD is merely a new method used to investigate tracer experiments and is found useful for the characterization of flow mal distribution. In the work of Robinson and Tester (1986), they proposed a new quantitative definition of sluggish flow or stagnancy through the use of the internal RTD. They also showed how the internal RTD and its related cumulative distribution function are used to describe the interior flow patterns. The internal cumulative RTD is defined:

11 57 ( 1 t 0 t 0 ( dt te( dt (4.7) The importance of the internal RTD is that it describes the eventual residence times, therefore the internal flow velocities of the different fluid elements and this is important for systems where a large degree of flow mal distribution occurs (Robinson and Tester 1986). In order to define the interior flow patterns within a system, the internal and external cumulative RTD functions, ( and F( respectively, are used. In the work of Robinson & Tester, they demonstrated the use of ( and F( by analyzing an idealized flow model, one of complete segregated flow Calculation The following formulas are utilized in the computation of internal and exit age distribution. 1. Experimental data: concentration ( C) vs time t. 2. Area under C vs t curve = t. c ( t t 3. Exit age distribution: E( C( C t 4. Exit age distribution (Dimensionless E( )): E( ) E( 5. Internal age distribution: I( ) = 1- E(

12 58 6. F-curve: E( ) = E(. Tables 4.1 to 4.5 represent the experimental Concentration Vs time and the computed values of and E( ). elements. Table 4.6 to 4.10 gives the results for annulus without mixing Table 4.1 Experimental RTD Studies in Annulus with Mixing Elements of Volumetric Flow Rate = 40 l/hr Time( time distribution E( )

13 59 Table 4.2 Experimental RTD Studies in Annulus with Mixing Elements of Volumetric Flow Rate = 60 l/hr Time( time distribution E( ) TABLE 4.3 Experimental RTD Studies in Annulus with Mixing Elements of Volumetric Flow Rate = 80 l/hr Time( time distribution E( )

14 60 Table 4.4 Experimental RTD Studies in Annulus with Mixing Elements of Volumetric Flow Rate = 100 l/hr Time( time distribution E( ) Table 4.5 Experimental RTD Studies in Annulus with Mixing Elements of Volumetric Flow Rate = 120 l/hr Time( time distribution E( )

15 61 Table 4.6 Experimental RTD Studies in Annulus without Mixing Element of Volumetric Flow Rate = 40 l/hr Time( time distribution E( ) Table 4.7 Experimental RTD Studies in Annulus without Mixing Element of Volumetric Flow Rate = 60 l/hr Time( time distribution E( )

16 62 Table 4.8 Experimental RTD Studies in Annulus without Mixing Element of Volumetric Flow Rate = 80 l/hr Time( time distribution E( ) Table 4.9 Experimental RTD Studies in Annulus without Mixing Element of Volumetric Flow Rate = 100 l/hr Time( time distribution E( )

17 63 Table 4.10 Experimental RTD Studies in Annulus without Mixing Element of Volumetric Flow Rate = 120 l/hr Time( time distribution E( ) CONCLUSION The solution of Silver nitrate solution (7 litres of 0.05 N) and 5N Sodium chloride is poured into the tubular reactor with static mixer. The same procedure is repeated for different flow rates. An identical annulus, without helical elements is also constructed to compare the performance of the static mixer. The procedures of model parameter RTD identification were proposed for the Lab- made, tubular reactor with static mixer. Experimental results of RTD were proposed for this tubular reactor.