Saturday, October 14, 2006

CENTRIFUGAL PUMPS APPLICATIONS

·Net Positive Suction Head for Pumps Handling Hydrocarbon Liquids and Water at Elevated Temperatures


The NPSH requirements of centrifugal pumps are normally determined on the basis of handling water at or near normal room temperatures. Operating experience in the field has indicated, and a limited number of carefully controlled laboratory tests have confirmed, that pumps handling certain hydrocarbon fluids, or water at significantly higher than room temperatures. will operate satisfactorily with less NPSH available than would be required for cold water.

Figure 61 is a composite chart of NPSH reductions which may be expected for hydrocarbon liquids and high temperature water based on available laboratory data from tests conducted on the fluids shown, plotted as a function of fluid temperature and vapor pressure at that temperature.

·Limitations for Use of Chart for Net Positive Suction Head Reductions (Fig. 61)

The following limitations and precautions should be observed in the use of Fig. 61.

Until specific experience has been gained with operation of pumps under conditions where this chart applies, NPSH reductions should be limited to 50% of the NPSH required by the pump for cold water.

This chart is based on pumps handling pure liquids. Where entrained air or other noncondensable gases are present in a liquid, pump performance may be adversely affected even with normal NPSH available (see below) and would suffer further with reductions in NPSH. Where dissolved air or other noncondensables are present, and where the absolute pressure at the pump inlet would be low enough to release such noncondensables from solution, the NPSH required may have to be increased above that required for cold water to avoid deterioration of pump performance due to such release.

For hydrocarbon mixtures, vapor pressure versus temperature relationships may vary significantly with temperature, and specific vapor pressure determinations should be made for actual pumping temperatures.

In the use of the chart for high temperature liquids, and particularly with water, due consideration must be given to the susceptibility of the suction system to transient changes in temperature and absolute pressure, which might necessitate provision of a margin of safety of NPSH far exceeding the reduction otherwise available for steady state operation.

Because of the absence of available data demonstrating NPSH reductions greater than ten feet, the chart has been limited to that extent and extrapolation beyond that limit is not recommended.

·Instruction for Using Chart for Net Positive Suction Head Reductions (Fig. 61)

Enter Figure 61 at the bottom of the chart with pumping temperature in degrees F and proceed vertically upward to the vapor pressure in psia. From this point follow along or parallel to the sloping lines to the right side of the chart, where the NPSH reductions in feet of liquid may be read on the scale provided. If this value is greater than one half of the NPSH required on cold water, deduct one half of the cold water NPSH to obtain corrected NPSH required. If the value read on the chart is less than one half of the cold water NPSH, deduct this chart value from the cold water NPSH to obtain corrected NPSH required.

Example: A pump that has been selected for a given capacity and head requires a minimum of 16 feet NPSH to pump that capacity when handling cold water. In this case the pump is to handle propane at 55 F, which has a vapor pressure of 100 psia. Following the procedure indicated above, the chart yields an NPSH reduction of 9.5 feet, which is greater than one half of the cold water NPSH. The corrected value of NPSH required is therefore one half the cold water NPSH or 8 feet.

Example: The pump of example above has also been selected for another application to handle propane at 14 F, where it has a vapor pressure of 50 psia. In this case, the chart shows an NPSH reduction of 6 feet, which is less than one half the cold water NPSH. The corrected value of NPSH is therefore 16 feet less 6 feet, or 10 feet.

·Use of Chart for Net Positive Suction Head Reductions (Fig. 61) for Liquids Other Than Hydrocarbons or Water.

The consistency of results which have been obtained on tests which have been conducted with both water and hydrocarbon fluids suggests that NPSH required by a centrifugal pump may be reduced when handling any liquid having relatively high vapor pressure at pumping temperature. However, since available data are limited to the liquids for which temperature and vapor pressure relationships are shown on Figure 6.1, application of this chart to liquids other than hydrocarbons and water is not recommended except where it is understood that such usage can be accepted on an experimental basis.

·Centrifugal Pumps Handling Entrained Air or Gas

Under a number of different circumstances, centrifugal pumps may be required to handle a mixture of air and water or similar mixtures. It is known that this reduces the head-capacity and efficiency of a centrifugal pump, even when relatively small percentages of air or gas are present.

Deterioration of performance for a given percentage of air or gas varies from pump to pump depending on rotating speed, specific speed, pump size, suction pressure, discharge pressure, number of stages and various special design features. These mixtures may also have a detrimental effect on the mechanical operation of the pump. An explanation and evaluation of the effect of these factors is beyond the scope of this article.

·Determination of Pump Performance When Handling Viscous Liquids

The performance of centrifugal pumps is affected when handling viscous liquids. A marked increase in brake horsepower, a reduction in head, and some reduction in capacity occur with moderate and high viscosities.

Figs. 62 and 63 provide a means of determining the performance of a conventional centrifugal pump handling a viscous liquid when its performance on water is known. They can also be used as an aid in selecting a pump for a given application. The values shown in Fig. 62 are averaged from tests of conventional single stage pumps of 2-inch to 8-inch size, handling petroleum oils. The values shown in Fig. 63 were prepared from other tests on several smaller pumps (1" and below). The correction curves are, therefore, not exact for any particular pump.

When accurate information is essential, performance tests should be conducted with the particular viscous liquid to be handled.

·Limitations on Use of Viscous liquid Performance Correction Chart

Reference is made to Fig. 62 and Fig. 63. Since these charts are based on empirical rather than theoretical considerations, extrapolation beyond the limits shown would go outside the experience range which these charts cover and is not recommended.

Use only for pumps of conventional hydraulic design, in the normal operating range, with open or closed impellers. Do not use for mixed flow or axial flow pumps, or for pumps of special hydraulic design for either viscous or non-uniform liquids.

Use only where adequate NPSH is available in order to avoid the effect of cavitation.

Use only on Newtonian (uniform) liquids. Gels, slurries, paper stock and other non-uniform liquids may produce widely varying results, depending on the particular characteristics of the liquids.


·Symbols and Definitions Used in Determination of Pump Performance When Handling Viscous Liquids

These symbols and definitions are:

Qvis = Viscous capacity in gpm
The capacity when pumping a viscous liquid

Hvis = Viscous head in feet
The head when pumping a viscous liquid

Evis = Viscous efficiency in percent
The efficiency when pumping a viscous liquid.

bhpvis = Viscous brake horsepower
The horsepower required by the pump for the viscous conditions

Qw = Water capacity in gpm
The capacity when pumping water

Hw = Water head in feet
The head when pumping water

Ew = Water efficiency in per cent
The efficiency when pumping water

spgr = Specific gravity

CQ = Capacity correction factor

CH = Head correction factor

CE = Efficiency correction factor

1.0 Qw = Water capacity at which maximum efficiency is obtained.

The following equations are used for determining the viscous performance when the water performance of the pump is known:

Qvis = CQ x Qw

Hvis = CH x Hw

Evis = CE x Ew

bhpvis =[Qvis x Hvis x Spgr] / [3960 x Evis]

CQ, CH and CE are determined from Fig. 62 and Fig. 63 which are based on water performance. Fig. 62 is to be used for small pumps having capacity at best efficiency point of less than 1.00 GPM (water performance).

The following equations are used for approximating the water performance when the desired viscous capacity and head are given and the values of CQ and CH must be estimated from Fig. 62 or 63 using Qvis and Hvis, as:

Qw (approx.) = Qvis/CQ

Hw (approx.) = Hvis/CH

·Instructions for Preliminary Selection of a Pump for a Given Head-Capacity-Viscosity Condition

Given the desired capacity and head of the viscous liquid to be pumped, and the viscosity and specific gravity at the pumping temperature, Figs. 62 or 63 can be used to find approximate equivalent capacity and head when pumping water.

Enter appropriate chart at the bottom with the desired viscous capacity, (Qvis) and proceed upward to the desired viscous head (Hvis) in feet of liquid. For multistage pumps use head per stage. Proceed horizontally (either left or right) to the fluid viscosity, and then go upward to the correction curves. Divide the viscous capacity (Qvis) by the capacity correction factor (CQ) to get the approximate equivalent water capacity (Qw approximately). Divide the viscous head (Hvis) by the head correction factor (CH) from the curve-marked "1.0 x Qw" to get the approximate equivalent water head (Hw approximately). Using this new equivalent water head-capacity point, select a pump in the usual manner. The viscous efficiency and the viscous brake horsepower may then by calculated.

This procedure is approximate as the scales for capacity and head on the lower half of Fig. 62 or Fig. 63 are based on the water performance. However, the procedure has sufficient accuracy for most pump selection purposes. Where the corrections are appreciable, it is desirable to check the selection by the method described below.

EXAMPLE: Select a pump to deliver 750 gpm at 100 feet total head of a liquid having a viscosity of 1000 SSU and a specific gravity of 0.90 at the pumping temperature.

Enter the chart (Fig. 63) with 750 gpm, go up to 100 feet head, over to 1000 SSU, and then up to the correction factors:

CQ = 0.95

CH = 0.92 (for 1.0 Qnw)

CE = 0.635

Qw =[750/0.95] = 790 gpm

Hw = [100/0.92] = 108.8 (~ 109 feet head)

Select a pump for a water capacity of 790 gpm at 109 feet head. The selection should be at or close to the maximum efficiency point for water performance. If the pump selected has an .efficiency on water of 81 per cent at 790 gpm, then the efficiency for the viscous liquid will be as follows:

Evis = 0.635 x 81% = 51.5 per cent

The brake horsepower for pumping the viscous liquid will be:

bhpvis = [750 x 100 x 0.90]/[3960 x 0.515] = 33.1 hp

For performance curves of the pump selected, correct the water performance as discussed below.

·Instructions for Determining Pump Performance on a Viscous liquid When Performance on Water is Known

Given the complete performance characteristics of a pump handling water. determine the performance when pumping a liquid for a specified viscosity.

From the efficiency curve, locate the water capacity (1.0 x Qw) at which maximum efficiency is obtained.

From this capacity, determine the capacities (0.6 x Qw), (0.8 x Qw) and (1.2 x Qw).

Enter the chart at the bottom with the capacity at best efficiency (1.0 x Qw), go upward to the head developed (in one stage) (Hw) at this capacity, then horizontally (either left or right) to the desired viscosity, and then proceed upward to the various correction curves.

Read the values of CE and CQ and of CH for all four capacities.

Multiply each head by its corresponding head correction factor to obtain the corrected heads. Multiply each efficiency value by CE to obtain the corrected efficiency values which apply at the corresponding corrected capacities.

Plot corrected head and corrected efficiency against corrected capacity. Draw smooth curves through these points. The head at shut-off can be taken as approximately the same as that for water.

Calculate the viscous brake horsepower (bhpvis) from the formula given above.

Plot these points and draw a smooth curve through them which should be similar to and approximately parallel to the brake horsepower (bhp) curve for water.

EXAMPLE: Given the performance of. a pump (Fig. 64) obtained by test on water, plot the performance of this pump when handling oil with a specific gravity of 0.90 and a viscosity of 1000 SSU at pumping temperature.

On the performance curve (Fig. 64) locate the best efficiency point which determines Qnw. In this example it is 750 gpm. Tabu1ate capacity, head and efficiency for (0.6 x 750), (0.8 x 750) and (1.2 x 750).

Using 750 gpm, 100 feet head and 1000 SSU, enter the chart and determine the correction factors. These are tabulated in Table of Sample Calculations. Multiply each value of head, capacity and efficiency by its correction factor to get the corrected values. Using the corrected values and the specific gravity, calculate brake horsepower. These calculations are shown below. Calculated points are plotted in Fig. 64 and corrected performance is represented by dashed curves.

Fig. 62 is used in the same manner as fig. 63 except that only one point on the corrected performance curve is obtained. Through the corrected head-capacity point, draw a curve similar in shape to the curve for water performance and having the same head at shut-off. If the capacity correction CQ is less than 0.050, the corrected head-capacity curve should be a straight line. The corrected-efficiency point represents the peak of the corrected efficiency curve, which is similar in shape to that for water. The corrected brake horsepower curves are generally parallel to that for water.

Radial Thrust in Single Volute Pumps Single volute pump casings in the specific speed range between 500 and 3500 may be designed for uniform pressure around the volute casing at the design or best efficiency point capacity. For pumps in applications normally operating at or near the best efficiency point capac.ity, the thrust factor may approach zero. On either side of the best efficiency point capacity, pressure distripution is not necessarily constant, resulting in radial thrust. The radial thrust at shutoff may be approximated, for this type of design, using the following expression:

Rso = Kso x [(Hso x spgr)/2.31] x D2 x B2

Thrust values R at capacities other than shut- off may be approximated by the following formula:

R = (K/Kso) x (H/Hso) x Rso

where K = Kso x [1-(Q/Qn)x]

and
Rso = Radial thrust in pounds, at shut-off

R = Radial thrust in pounds, at operating condition

Kso = Thrust factor at shutoff from Fig. 65

K = Thrust factor at operating condition

Hso = Total head at shutoff in feet

H = Total head at operating condition in feet

spgr = Specific gravity of the liquid

D2 = Impeller diameter in inches

B2 = Impeller width at discharge, including shrouds in inches

Q = Capacity at operating condition, in gpm

Qn = Capacity at best efficiency point in gpm

x = Exponent, varying between 0.7 and 3.3 established by test. In the absence of test data, the exponent may generally be assumed to vary linearly between 0.7 at specific speed 500 and 33 at specific speed 3500.




Saturday, September 16, 2006

Industrial Noise: Properties, Sources and Solutions

Presented is an overview of the sources of industrial noise, the fundamentals of sound and how noise levels can be reduced

ANY EFFECTIVE effort to lessen noise pollution dictates a familiarity with its most basic constituent - sound. How sound is generated in industrial machinery and how sound can be inhibited are the subjects of this article.

The sound a person hears can be either pleasant or annoying depending upon the volume and frequency of that sound. Annoying sounds most commonly are referred to as noise. Accurate measurement of sound is very important in the analysis and abatement of noise.

Sensitivity of the human ear varies with the frequency and pressure level of the sound. Therefore, in measuring and analyzing noise, a weighting system must be employed. Originally, three weighting scales were established such that for low volume sound scale "A" was used. Scale "B " was used on medium volume sound and scale "C" on high volume sound. The "A" scale., however, has emerged as most closely approximating the ear and is the most commonly used weighting scale.

Measuring instruments consist of a microphone, amplifiers, weighting circuits and an output meter. To get an accurate reading to determine the noise level of a given source, background. noise must be considered. Anechoic (sound absorbing) or more often semi-anechoic rooms are used to eliminate background and reflected noise.

SOURCES OF MACHINERY NOISE

Sound is generated whenever there is sufficient disturbance in a medium to cause a detectable pressure fluctuation. In the operation of industrial machinery where moving parts and fluid flow are involved, there are numerous disturbances which may be considered noise sources.

Gear noise is related directly to the design, manufacturing technique and operation of the gears. Friction associated with the rolling and sliding of metal to metal contact causes vibration and noise. Accuracy of tooth form and spacing is necessary to reduce the small changes in acceleration and impacts that cause the vibration and excitation resulting in excess noise.

Although bearing noise generally is small compared to other noise sources, it still deserves consideration. In journal bearings, noise is caused by relative motion between the shaft and the bearing surface. Inadequate oil supply and irregularities in shaft or bearing surfaces tend to increase noise. With ball bearings, the ball movement between the inner and outer races generates noise. Again, imperfections in the surfaces and inadequate oil supply can cause excess noise. Tilt pad type bearings have mechanical movement of the pads that can add to the noise generated.

Noise generated by couplings is due mostly to windage and can be greatly reduced by proper design of the coupling guard.

Noise caused by a mechanical imbalance generally is very small. However, noise generated from the mechanical or hydraulic reaction associated with this imbalance can be quite noticeable. Mechanical or hydraulic forces caused by imbalance are proportional to the square of the speed. Therefore, the effect of an imbalance becomes of greater concern at higher speeds.

Sound produced by a machine normally is related to input horsepower. Actual increase in noise associated with a given increase in power can be determined only by test data. However, a general rule of thumb for centrifugal pumps and compressors is that the sound pressure level will slightly more than double with a doubling of input horsepower (approximately 3 to 4 dbA).

In general, high speed machines seem noisier than low speed machines. However, there is no simple formula which relates rotational speed to noise. The increase in noise generated is a function of speed, the ratio of rotating mass to the base, type of machine, mounting configuration, machine alignment, excitation frequencies and structural resonance. Test data is the best way to establish the effect of rotational speed on a given machine design.

In addition to mechanical noise, pipe flow noise presents problems as well.

The major cause of noise in piping systems is related directly to turbulent flow. In most every industrial piping system, turbulent flow is present. Turbulent flow is caused by various items such as high velocity, changes in pipe diameters, bends, restrictions and obstructions in the flow path.

This piping noise may well exceed the noise radiated from the machinery installed in the system. Therefore, it is important to consider reducing the pipe noise as well as the machinery noise.

Aerodynamic noise in compressors. Noise associated with the flow of gas through a system is referred to as aerodynamic noise. In centrifugal compressors this aerodynamic noise is generated mostly by turbulence in the flow path. Centrifugal compressors work on the basis of increasing gas velocity then diffusing the velocity head to pressure head. This, of course, requires high velocities and, therefore, considerable turbulence. This turbulence also is increased by Von Karman vortices which follow the trailing edge of the impeller blades.

If flow into the eye of the impeller is turbulent, this will tend to increase the noise as well. Therefore, it is important to have smooth, well finished flow passages in the case and diffuser. Obstructions and abrupt t changes in passage diameters also should be avoided.

When a compressor operates in a surge condition, a hunting situation is established and flow reversal occurs. This, of course, increases turbulence and flow noise. Since there is also a considerable amount of vibration associated with compressor surge, this will increase the mechanical noise as well.

The terms choke or stonewall describe the maximum flow condition in the compressor. In this situation, the relative condition of sonic flow occurs in the compressor diffuser. This generally results in an increase in the noise due to the increase in turbulent mixing downstream of the sonic flow point and also due to the shock mechanism. Both surge and choke situations should be avoided not only for noise reduction but also to prevent equipment damage and maintain system performance.

Hydraulic noise. Noise generated by turbulent flow of liquid in a pump is similar to aerodynamic noise in compressors. High speeds associated with the centrifugal impeller and vortices created by the impeller blades result in added turbulence. It again is necessary to maintain smooth, unobstructed flow passages to keep turbulence to a minimum.

In both pumps and compressors, increased turbulence is associated with a decrease in unit efficiency. Operation at or near the best efficiency point will tend to reduce hydraulic noise.

Pump cavitation is a situation where, due to a local pressure drop, cavities filled with vapors are formed and then collapse at higher pressures than the liquid vapor pressure. In this situation, the cavity collapse locally exerts forces on the pump. This not only in- creases turbulence and flow noise but increases mechanical noise and can cause damage to the equipment.

Since viscosity of a liquid is greater than that of a gas, hydraulic turbulence associated with a given fluid velocity will be less. Hydraulic noise, therefore, is less prevalent than its counterpart, aerodynamic noise.

Electric motor noise. Electric motor noise is complex and a combination of the following:

· Windage noise due to cooling air flow turbulence .
· A siren or whistling noise produced by the fan blades passing close to stationary members
· Rotor-slot noise caused by rotating open slots
· Noise generated by combination of rotor bars and stator slots
· Noise due to high flux density
· Bearing and other mechanical noise such as misalignment and dynamic unbalance
· Unbalanced line currents in three-phase power supply can increase noise produced by the motor.

INHIBITING SOUND

Now that the many sources of machinery noise have been delineated, what methods of sound control can be employed?

The best means of controlling sound is to prevent noise generation or at least reduce its level. In some cases this is possible, but generally, machinery is designed with performance in mind, and to change internal components to reduce noise also can reduce efficiency. Therefore, it is necessary to find other ways to reduce the noise levels that reach the listener.

The basic methods of reducing noise are:

Sound absorption. Machinery installed in enclosed rooms has the added problem of noise that reverberates off the walls, thus increasing the noise level the operator will hear. This reverberant noise can be reduced substantially by lining the. walls of the room with a soft porous material that will absorb the sound and reduce sound reflected back into the room.

Sound isolation. A common way of reducing machinery noise is to place a sound barrier between the sources and the listener. Effectiveness of this barrier is described by its transmission coefficient, which is defined as the fraction of the incident sound transmitted through the barrier.

Sound is transmitted through a solid barrier by forced vibration of the wall caused by sound wave striking the surface. The heavier, more rigid, and more airtight the barrier is, the more resistant it is to sound transmission. Internal dampening and bending stiffness effect sound transmission as well.

Vibration isolation. Airborne noise generated by a vibrating part generally can be reduced by isolating this part from the rest of the structure or machine. The simplest isolating device is a flexible support that reduces the magnitude of the: force that would be transmitted to the structure or machine. By the same token, an isolator may reduce the amplitude transmitted from a vibrating support to a part of the machine generating excess noise.

Vibration dampening. Vibrating parts have certain resonant frequencies. When an exciting force has the same frequency as the resonant frequency of that part, the amplitude will be limited only by the amount of dampening in the system. Noise generated by this resonant part can be reduced by increasing the dampening in the system.

In the absence of all dampening, the amplitude of a vibrating part will go to infinity. Of course, there is always some dampening in all systems. Adjustment of this dampening is one of the most important factors in vibration and noise control.

Mufflers. Silencers or mufflers generally are divided into two categories: absorptive and reactive. The distinction between these two is somewhat arbitrary since nearly all mufflers use a combination of both to accomplish noise reduction.

Absorptive mufflers rely on the dissipation or absorption of the sound into sound-absorbing material. Internally-lined ducts, plenum chambers and baffle mufflers fall in this category. They often are used with fans, jet engines, gas turbines, and air ejectors.

Reactive mufflers utilize the reflective properties of sound and do not rely on sound absorption. Conical connectors, expansion chambers, resonators and tail pipes are used in these mufflers to accomplish noise reduction. The most common of these is the automobile muffler.

Acoustic enclosures. When a large amount of noise reduction is desired, an acoustic enclosure is often the most direct solution. The enclosure serves as a noise barrier that completely encloses the machine. It is de- signed such that its resonant frequency and the noise frequency of concern do not coincide.

Enclosures with hard, reflective walls may have an internal sound pressure level higher than if no enclosure was present. This can be avoided by lining the enclosure with a sound absorptive material.

The number and size of the openings in these enclosures are of concern as they can greatly reduce its effectiveness. Of course, each individual plant situation will differ in types of noise and possible control methods. As a practical matter, it may not be possible to reduce noise as much as we would like. Nevertheless, by being aware of the sources of sound pollution and possible alternatives it is possible, in most cases, to reduce greatly the problem of industrial noise pollution and make the work place safer and more productive for worker and manager alike.


Plant Energy Conservation

An overview of options for reducing energy use in equipment, processes and utilities

Opportunities for energy conservation in processing plants exists primarily in three areas: equipment, unit operations and processes, and utilities. Fired heaters, heat exchangers and shaft work systems are part of the equipment area. Separation and reaction systems are part of the unit operations and processes area. The utilities area includes the steam/power system, .cooling system and fuel system. Suggestions are provided in each of these areas to help focus efforts on proven techniques for reducing plant energy consumption.

EQUIPMENT

Fired heaters. Methods to improve the efficiency of fired heaters focus on minimizing the loss of useful heat to the surroundings from the hot flue gases. This can be done in. two ways. The quantity of the flue gas can be reduced (within limits set by the stoichiometric requirements for combustion) and the temperature of the flue gas can be reduced by recovering more heat. The quantity of air supplied is set by the oxygen requirements for complete combustion of the fuel. Practically, a slight excess in air is required to ensure complete combustion since fuel and air mixing at the burner is never totally complete. At low excess air, losses result from unburned fuel, while at high excess air the losses result from the heating and release of the excess air. The optimum operating point can be determined from measurements of the CO content in the flue gas. Excess air should be monitored and reduced until the CO concentration in the flue gas just begins to increase.

High flue gas temperatures can be reduced by recovering additional heat to preheat either the combustion air and/or a boiler's feed water. Air preheaters and economizers are usu ally used to recover heat from hot flue gases. The quantity of heat which can be recovered with either is set by the acid dew point of the flue gas. Flue gas energy recovery usually requires additional investment and is principally done during initial design since retrofit in the limited space around a process furnace or boiler is often difficult.

In a boiler it is possible to reduce the fuel required to make steam by raising the temperature of the boiler feed water. There are usually many low temperature hot process streams in a plant which are not hot enough to transfer heat to cold process streams economically but can be used as a source of heat for boiler feed water. This heat leads to a direct reduction in fuel use. Another method to reduce the fuel consumption in boilers is to flash the boiler blowdown to produce low pressure steam. The desirability of this will depend on the current plant requirements for low pressure steam.

Depending on the fuel burned, furnace and boiler tubes can become coated with deposits. This increase the resistance to heat transfer resulting in higher flue gas temperatures. To maximize the use of available heat transfer surface, soot blowing frequency and decoking of tubes should be optimized.

Heat Exchangers. Optimizing of heat recovery involves maximizing the transfer of heat from hot process streams which need to be cooled to cold streams which need to be heated. Targets for heat recovery can be obtained by using "pinch" technology. This technique considers what quantity of heat recovery from hot to cold process streams is possible if the streams could be broken into incremental pieces. Some advanced applications of the "pinch" technique also allow the preliminary identification of heat-exchanger services which are improperly placed and, therefore, may be operating inefficiently.

Efficiency of a heat recovery system normally decreases with time due to increased fouling of the heat exchange surface. Fouling can be reduced by periodic chemical or mechanical cleaning of the exchanger surface, and by the addition of antifoulants. The economic impact of fouling is considerable because in addition to the loss of heat recovery capability, there .are significant maintenance arid energy costs associated with the disassembly and cleaning of the heat exchangers. Therefore, there is a need to balance these costs and set an optimum exchanger cleaning schedule based on minimizing the total annual cost of all the fouling related expenses.

A whole new generation of heat exchange equipment is available. These new exchangers, including spiral and plate/frame, provide several advantages. In many applications, they are less prone to fouling and more easily cleaned. More importantly is the application of these exchangers to services where there is a very small temperature difference between the hot and cold process streams. Because of their configuration, they can effectively transfer heat with a much lower approach temperature than a shell/tube heat exchanger.

Where fouling is a recurring problem, it can be economical to clean the heat exchanger without disassembly. On-line mechanical cleaning of exchangers, once limited to water services, is now being applied to hydrocarbon services. These include brushes and balls which repeatedly pass through the heat exchanger tubes, cleaning the surfaces.

A high heat transfer coefficient is important in maximizing the efficiency of a heat recovery system. For highly viscous fluids, gas flows, or systems operating at turndown, the heat transfer coefficient is often the key limiting factor. In these cases, it may be possible to install turbulence promoters to increase the heat transfer coefficient for the limiting side.

Power providers. Substituting high efficiency motors when replacing standard motors, or as an alternative to the repair of existing motors can lead to immediate savings.

For power needs which consistently vary, or are well below capacity, there are two alternatives. The first is to use variable speed drivers to more closely track the requirements. For long-term turndown operation, impellers can be trimmed avoiding the need for pressure reduction or recycle in addition to reducing energy consumption.

TABLE 1 - Fired heaters/boilers
· Monitor CO and excess air to reduce rejected energy and improve efficiency.
· Consider installation of economizers and air preheaters to recover additional heat from flue gas.
· Preheat boiler feedwater with available low temperature process streams to reduce fuel consumption.
· Maximize use of heat transfer surface by optimizing sootblowing frequency and decoking of tubes.
· Flash blowdown to produce low pressure steam if required.

TABLE 2 - Heat exchangers
· Retrofit heat exchanger networks to maximize heat recovery from existing process streams.
· Use "pinch" techniques to set energy recovery targets and identify inefficient heat exchanger services.
· Determine optimum heat exchanger cleaning schedule using total fouling related expenses method.
· Use new generation heat exchangers such as plate-frame and spiral for increased heat transfer and closer approach temperatures.
· Consider "on-line" mechanical cleaning of exchangers where fouling is a problem.
· Use turbulence promoters in laminar flow and gas phase services and where turndown has significantly reduced the heat transfer coefficient.

TABLE 3 - Power providers
· Use high efficiency motors when replacing or repairing existing installations.
· Substitute variable speed drivers in services where operation is frequently below capacity.
· Trim impellers for continuous turndown operation. . Reduce recycle in compressor operations.
· Consider expanders on FCCU regenerator flue gas and other pressurized streams.

TABLE 4 - Separation systems
· Optimize heat integration between feed and product streams and also between towers in sequence.
· Maximize removal of heat at the highest temperature in pumparounds by use of mid-condensers.
· Optimize the use of the lowest level of heat through the use of mid-reboilers.
· Minimize tower pressure to reduce reboiler duty.
· Replace inefficient tower internals by installing packing and higher efficiency trays.
· Optimize feed location to take full advantage of entire tower.
· Consider rearranging light-ends system for more efficient operation.
· Replace steam jet-ejectors in vacuum pipestills with vacuum pumps.
· Minimize overfractionation when subsequent mixing occurs downstream.
· Replace overhead condensers with extended surface bundles to allow decreased tower temperature.
· Eliminate utility cooling on tower pumparounds.


Numerous processes operate at elevated pressures. Power can often be extracted from the product streams of these units. Typical of such power recovery systems is the use of an expander on a fluid cat cracker regenerator. Although this decreases the quantity of steam which can be generated in a CO boiler, it is more than offset by the additional power provided.

UNIT OPERATIONS AND PROCESSES

Separation systems. Separation systems otter many opportunities for heat integration due to the usually large number of product streams which need to be cooled. In addition, better efficiencies can be realized by improving the contacting.

The atmospheric and vacuum pipestills are among the refinery units which have been the subject of most energy conservation efforts. The principle focus of these efforts is to maximize the heat transfer from the hot product streams to the entering crude oil. This will result in a direct savings in furnace fuel firing. There are numerous techniques and computer programs available to optimize the heat recovery network. Many of these determine theoretical heat recovery targets, establish the optimum network configuration, and size the heat exchangers.

The most energy efficient fractionation tower would provide heat at the lowest temperature and remove it at the highest. It would have condensers for each tray above the feed and reboilers for each tray below the feed. In such a configuration, a higher level of heat can be extracted by some of the condensers while a lower level of heat can be provided to some of the reboilers. Unfortunately, such an arrangement of reboilers and condensers for each tray is not practica1. It is possible, however, to place mid-condensers and mid-reboilers at several locations.

Another way to reduce reboiler duty is to lower the tower operating pressure. Not only will this result in less required energy, but it will also lower the temperature at which the heat is required providing greater opportunity for heat integration.

Improving the contacting within the tower can effectively reduce the amount of energy used. Substituting packing for trays has long been practiced for sidestream strippers and vacuum pipestills. A whole new generation of "structured" tower packing is now available which can result in maximizing energy use by producing higher yields with the same amount of energy.

Quite often, parts of a tower are not used efficiently be- cause the feed is introduced on the wrong tray. When feed is incorrectly introduced, the useful tower height is significantly reduced due to the effect on the concentration gradient. Some analysis is necessary to correct this problem and locate the optimum feed location, but the improvement in yield and the potential reduction in energy use are well worth the effort.

One retrofit for a vacuum pipestill which can significantly reduce steam energy consumption involves replacing the steam jet-ejectors with vacuum pumps. This can be a significant steam saver.

In light-ends systems, several of the towers. are usually heat integrated. It is often possible to adjust tower pressure to facilitate this integration and maximize the heat recovery. In addition, it may be advantageous to consider alternate sequences for the separation. Such changes may increase the amount of heat recovery thus lowering the energy required for tower reboiling.

One way to reduce the amount of energy used for separations is to separate only what you need and do it only once. There are often several streams which are recovered separately from a fractionator only to be subsequently remixed. For example, two sidestreams from a vacuum pipestill may both become feed to a catalytic cracker. In this case, is it not worth using energy to ensure that product specification is achieved for each of these streams. It is not uncommon in a complex plant to have streams separated, remixed, and then partially separated again.

In many cases, decreasing the tower pressure is limited by the temperature of the cooling utility and the duty of the overhead condenser. This limitation often can be overcome by allowing a lower approach temperature in the condenser by using extended surface enhanced heat exchanger tubes. Bundle replacement with low-fin tubes is relatively inexpensive since new foundations or piping are not-required.

Finally, the amount of energy which is literally "thrown away" should be minimized. This includes any streams which are cooled by utilities at temperatures where useful heat recovery is potentially economic. Often there is no heat recovery from the atmospheric pipestill overhead even though this represents a large quantity of heat, albeit at a rather low temperature. It is possible to use this low level energy to start heating the entering crude or boiler feed water. Of greater concern is utility trim cooling on tower mid-condensers and pumparounds. This is usually higher level heat which can be used easily to reboil other fractionators or make steam.

Conversion systems. Many conversion systems, such as hydrotreaters and crackers, produce product streams at significantly higher temperatures than their feed stream. By recovering this heat, less fuel will be required to heat the feed up to the reactor inlet temperature.

Many conversion units are operating at reduced thruput due to excess capacity. One way to save energy in these situations is to decrease the operating severity while keeping the same conversion. This can result in significant energy savings from such things as longer times between catalyst regeneration, and lower recycle compression.

In some cases new, more energy efficient processes can be substituted for existing units. The new separator technology is an example. It consumes much less energy than other methods of recovering hydrogen and can be retrofit easily into a plant. Process changes are also possible, such as substitution of newer solvents for gas treating where 'improved operation as well as energy credits result.

Using higher concentration hydrogen also can result in energy savings in several processes. In hydrotreaters the total feed, including treat gas, needs to be heated. A higher hydrogen concentration reduces the amount of feed and thus the heating requirements.


UTILITIES

The utilities area consists of three energy systems: the steam/power system, the cooling system and the fuel system.

For steam systems, the system balance should be optimized using one of the many computer programs. These allow analysis of the trade-offs between ways of producing shaft work and providing steam heating, and can result in significant boiler fuel savings. One of the largest sources of energy losses in steam systems is from the steam traps. Individually the losses may be small, but there are hundreds of traps throughout a plant. A documented maintenance program should be established to repair or replace leaking or malfunctioning traps.

TABLE 5-Conversion systems
· Maximize feed/effluent heat recovery to reduce furnace fuel requirements.
· Decrease operating severity during turndown to reduce energy consumption at same conversion.
· Minimize recycle gas in reformers to save both compression energy and furnace fuel.
· Consider replacing H2 plant with separators.
· Increase concentration of hydrogen in gas feeds to hydrotreaters to reduce fuel for gas heating.
· Recover pressure energy from FCCU regenerators and cokers using gas expanders.
· Substitute newer solvents in amine gas treating units.

TABLE 6- -Utility systems
· Optimize steam/power balances using computer software.
· Establish steam trap maintenance program to reduce steam leaks.
· Consider electrical tracing of process lines where only part-time heating is required.
· Maximize cogeneration through use of backpressure steam as a heat source.
· Minimize peak electrical power demand since this is a major rate setting factor.
· Replace inefficient cooling tower internals to reduce available temperature of cooling water.
· Maximize efficiency of cooling utility by allocating coldest water to compressor suctions and tower overheads.
· Maximize sale of components from fuel gas to allow for additional energy conservation measures.


In many cases, tracing of process lines is only necessary at night or in cold weather. Steam tracing systems often are not able to respond quickly to changing ambient conditions and are left on all of the time resulting in wasted energy. An alternative is to install electrical tracing which can be controlled much more easily with changes in ambient conditions.

The highest energy efficiency for steam results when it is used more than once. Using backpressure turbines, not only can power be extracted, but the rejected steam can be used for lower level process heating. This cogeneration of power and heat should be maximized to reduce total energy costs.

The cost of electrical power is often set based on peak demand. Minimizing this by using standby gas turbines often can result in a significant savings in overall plant power costs.

Cooling towers can be upgraded by replacing the internals which can improve contacting and result in lower temperatures. Proper allocations of the coldest water to compressor suctions and tower overheads optimizes use of the utility.

The fuel system balance can have a major effect on energy conservation efforts. Savings which result in additional gas flaring are not savings at all. The energy value of the fuel is still expended. Attempts to maximize the recovery of salable products from the plant gas system can help to alleviate this "gas containment" problem.

Tables 1 through 6 provide a checklist of energy saving opportunities in the areas discussed.

Look at Claus Unit Design

Many factors are fundamental to good design of the Claus unit

ELEMENTAL SULFUR is produced by reacting H2S and S02 in the vapor phase.

2H2S + SO2 --> 1.5S2 + 2H2O

Reaction products are cooled to about 290°F to condense which is subsequently separated in the molten state from uncondensed gases. Sulfur is generally stored and/or shipped as a liquid at temperature above 250°F or in various solid forms. Liquid sulfur has a clear, bright cherry red appearance while solid sulfur varies in color from canary yellow to dark brown or black.

Claus sulfur recovery units consist of six primary operations:
· Combustion step
· Gas cooling
· Reaction of H2S and S02
· Condensation of elemental sulfur
· Reheating reactor feeds
· Incineration.

Sulfur is produced both non-catalytically (combustion zone) and catalytically to reach maximum recovery. Generally one thermal reactor (combustion furnace) is followed in a series by two or more catalytic reactors. The ratio of reactants is controlled by fixing the amount to be burned.

H2S + 1.5O2 = SO2 + H2O

Since two moles of H2S react with one mole of SO2, only about one-third of the H2S feed is burned, leaving two-thirds to react with the newly formed SO2.

Because of the cooling required to condense elemental sulfur, gases need to be reheated prior to being fed to the downstream catalytic reactor. Following the final condensation and separation of molten sulfur, tail gases are usually sent to an incinerator or a tail gas treater. In the incinerator all unrecovered sulfur compounds are converted to SO2 prior to emission to the atmosphere. In a tail gas treater most of the unrecovered sulfur compounds are converted, and the remainder are usually incinerated before discharge to the air.


COMBUSTION

The combustion process is very critical to successful Claus operations. About one-third of the H2S is burned along with other feed combustibles: hydrogen, hydrocarbons, carbon monoxide, ammonia, cyanides, mercaptans, etc., which are partially oxidized. The furnace operates in a reducing atmosphere, which makes air feed control to the burner very important. Generally, only enough air is fed to oxidize about one-third of the H2S to SO2 and H2O, reduce hydrocarbons only to CO and H2O, convert NH3 to N2 and H2O, and burn cyanides and mercaptans to CO, N2, SO2 and H2O.

Normally, less than 40% of total feed gas is burned in the combustion chamber. For a typical Claus unit feeding a gas containing 81 % H2S, 9% CO2, 6% H2O and 4% hydrocarbons, only 31 % of the entire feed stream is oxidized. One- third of the H2S is burned (27 % of the feed) plus the 4% other combustibles. Thus combustion products include two-third of feed H2S plus some CO from partial oxidation of hydrocarbons. Additionally, some feed H2S converts directly to H2 and sulfur, creating a highly reducing atmosphere since there is no uncombined oxygen and there are significant amounts of reducing gases (H2S, H2, CO). Combustion air is usually supplied by a blower, since the furnace operates under a positive pressure (3.0 to 8.0 psig).

The combustion process is complicated considerably by the presence of ammonia, other mercaptans and cyanides in the feed. These compounds are very difficult to burn in a reducing atmosphere. Unoxidized cyanides and ammonia react with available H2S to produce compounds that plug the catalytic reactors, sulfur condensers, and heat exchangers, molten sulfur drain legs, etc. Therefore, it is imperative that ammonia or cyanide in the feed be properly destroyed by combustion.

A typical low H2S feed might contain 21 % H2S, 70% CO2, 6% H2O and 3% other combustibles. Here only 10% of the entire feed stream would be oxidized in the furnace (one-third of the feed H2S (7%) plus 3% other combustibles). This makes it difficult to maintain stable combustion in the furnace. Therefore, a procedure has been developed where some feed is bypassed around the furnace. This practice is called "split-flow" operation.

A potential disadvantage of "split-flow" is that some hydrocarbon, ammonia, cyanides, etc., are also fed directly to the first catalytic reactor without being burned. Unoxidized NH3 and cyanides tend to react with H2S to form compounds that plug the Claus unit. Often unreacted hydrocarbons degrade to carbon, shortening catalyst life. Carbon contamination can cause frequent unit shutdowns, premature catalyst change-outs, and off-color sulfur product.

Another technique developed to improve combustion stability involves use of preheated air and/or preheated acid gas. Typically, air and acid gas feed temperatures are less than 200°F and 100-150°F, respectively. By externally preheating air and/or acid gas to the burner, a more stable combustion condition is possible.

Theoretically, there is no upper limit to the air temperature. Practically, however, air is not preheated above 750°F, and more often is around 450°F. On the other hand, acid gas should not be heated above 650°F, since H2S starts to sulfide carbon steel at this temperature, damaging or destroying exchangers, piping, etc. Generally, acid gas is preheated to a maximum of 450°F.

Another procedure used for extremely low H2S feeds (<8%) is "direct oxidation." However, some variation of preheated feed and/or "split-flow" is usually preferable. Other approaches include addition of fuel to the unit feed, use of oxygen or oxygen-enriched air. However, use of oxygen usually is practical only if an air separation unit is part of the plant facilities.

Based on use of air for combustion, the types of Claus plants commonly selected are listed in Table 1.

TABLE 1

H2S in feed, % Type Claus unit
55-100 Straight through
40-55 Straight through with feed and/or air preheat
25-40 Split flow
12-25 Split flow with feed and/or air preheat
7-12 Split flow with feed and/or air preheat with added fuel
< 7 Claus plant usually not practical


For oxygen or oxygen-enriched air supply, these feed H2S concentrations would change somewhat.

All solvents used in sour gas treating result in hydrocarbons being present in Claus plant feed. Usually the hydrocarbon level is a minimum of 2% with a maximum of about 10%. Hydrocarbon content of Claus plant feed is a function of feed gas pressure, type of solvent used, and type of hydrocarbon present.


Gas Cooling & Sulfur Condensation

Gases leaving the combustion furnace at 2,000 to 2,500°F are most often cooled by generating saturated steam in a waste heat boiler of the fire-tube type. Hot gases pass: through the tubes with boiling water on the shell side where steam pressures between 20 and 600 psig can be successfully maintained. Sulfur is not normally condensed m the waste heat boiler because the outlet temperature (500 to 650°F) is I above the sulfur dew point temperature for gases leaving the furnace.

For straight-through type units, furnace-produced sulfur is roughly 50-65% of total production. The first sulfur condenser downstream of the waste heat boiler recovers sulfur by cooling gases to 300-400°F. Cooling is usually accomplished by steam generation at steam pressures between 20 and 80 psig. Gas outlet temperature from the sulfur condensers is limited to about 20oF minimum above steam temperature.

If low-pressure steam (20-80 psig) is not needed in the plant, another procedure occasionally used is a "closed" water boiling system where steam is generated in sulfur condensation at 20 psig, condensed and the condensate returned to the "boiler" (sulfur condenser) as feed water, thereby "closing the loop." This cooling method minimizes the amount of make-up water required for the system,


REACTOR FEED HEATING

Gas leaving the sulfur condenser is at its dew point but generally contains some liquid entrainment as well. Therefore, liquid sulfur needs to be vaporized and the gas temperature raised above the sulfur dew point before this gas can be fed to the catalytic reactor. Otherwise liquid sulfur would plug-off individual catalyst pores, effectively deactivating the catalyst. Additionally, elemental sulfur could solidify to form a concrete-like mass with the catalyst and block off gas flow through the bed. Therefore some type reactor feed heater is required following each sulfur condensation step except the final one.

Several reheat methods commonly used include:
· Indirect
· Fired heaters
· Hot gas bypass (direct mixing).

Direct mixing usually feeds a slip-stream of hot gas (600 to 1,500°F) from the waste heat boiler to the sulfur condenser outlet gases, upstream of the catalytic reactor. Mixing these two streams establishes the desired reactor inlet temperature.

Fired heater types include in-line burners wherein either fuel gas or acid gas is burned and combustion products mixed directly with reactor feed to raise the gas temperature to the proper level. Most in-line burners are fueled by a portion of the Claus unit acid gas feed.

Indirect reheat methods include use of steam or hot oil to heat reactor feed gases, which flow through the tube side of an exchanger upstream of the reactor. Also hot gases from a reactor outlet or boiler can be used as the heating media for the reactor feed, using a gas-to-gas heat exchanger.

The main advantage for hot gas bypass is its relatively low installation cost. This results since only piping and valves are required and the pressure drop is low. The main disadvantage is lower overall net sulfur recovery. Boiler outlet gases used as heat source normally contain large quantities of uncondensed elemental sulfur which bypass the first sulfur condenser (and often the second) and are fed directly to the ) catalytic reactor. This elemental sulfur in reactor feed limits the net reaction of H2S and SO2 because of the effects on re- action constants. Since the Claus reaction is reversible, there could be sufficient elemental sulfur to prevent net reaction of H2S and SO2 (2H2S + SO2 <=> 1.5 S2 + 2 H2O). This disadvantage is amplified at low turndown rates.

Because of the elemental sulfur problem, hot gas bypass method is usually limited to the first two catalytic reactors. Units with three catalytic reactors should not use this method on the third reactor. Usually an indirect method of reheat is preferable for the third reactor.

The main advantage of in-line burner reheat is ability to heat reactor catalyst to temperature levels where catalyst can be rejuvenated or regenerated. System pressure drop is also relatively low. Normally a portion of the acid gas feed is by-passed around the main combustion furnace and fed directly to the in-line reheater burner. Consequently, each in-line heater must be properly controlled to maintain the optimum H2S/SO2 ratio in the catalytic reactor feed for maximum sulfur production. Instead of controlling the air/acid gas ratio at only the main burner, this method requires control of air feed at four locations on a three catalytic reactor plant. Additionally, all acid gas does not pass through the thermal reactor where the largest amount of sulfur is produced. Some of acid gas also bypasses the first catalytic reactor, which generally produces about 30% of the total overall elemental sulfur. While this method generally results in better sulfur recovery than hot gas bypass, sulfur recovery is somewhat less than with the indirect methods.

Another potential disadvantage of in-line burner design is formation of SO3 in the system if air control is inaccurate. These sulfates can rapidly deactivate catalyst and, if the air rate is too low, produce carbon from hydrocarbon in the feed. Carbon is thus fed directly to the catalytic reactor, which can cause a premature catalyst change-out.

The most commonly used indirect reheat method is use of a steam-heated exchanger installed between each sulfur condenser and catalytic reactor. Process gas passes through tubes with condensing steam on the shell side. This is the most accurate control method and results in the best overall sulfur recovery performance. Also there is no loss of recovery efficiency at reduced unit thruput rates. The main disadvantages of this method are that pressure drop across these exchangers is higher than in other systems and the initial investment is more. Additionally, reactor inlet temperatures are limited to some level below that of the heating media. Thus catalyst regeneration is usually not possible with indirect reheat. However, indirect reheat avoids catalyst sulfation, carbon deposits, sulfur laydown, etc., thus virtually eliminating need for regeneration. Therefore, indirect reheat actually results in longer catalyst life.


INCINERATION

Tail gas is usually incinerated, since some sulfur always passes through the Claus process. Incineration oxidizes all sulfur compounds to SO2 such that gases can be discharged to the atmosphere via a stack. Stack height is theoretically a function of the amount and/or concentration of SO2 in the gases, although it is usually governed by local pollution regulations. Where no local or state regulations set the height, a safe discharge of SO2 can usually be made at about 100 feet, above grade.

Tail gas also can be fed to a downstream tail gas treater designed to recover most the sulfur in the stream. However, off-gas from the tail gas treater is frequently incinerated for the same reason, i.e., to convert sulfur compounds to SO2 before discharge to the atmosphere.


OPERATING GUIDES

Minimum temperature for stable combustion of the main acid gas feed stream is about 1,750oF. For better results this temperature should exceed 1,850°F, and preferably should be above 2,000°F. Maximum temperature usually is about 2,600°F.

Sulfur condenser outlet temperatures typically are 350- 400°F for the first two condensers and 270-320°F for the final two condensers. Lower sulfur condenser outlet temperatures result in higher overall sulfur recovery.

The approach to equilibrium for the Claus reaction is better realized at as low a temperature as possible. Conversely, carbon-sulfur compounds such as COS and CS2, which are formed in the combustion operation (Table 2), react more readily at higher temperatures (Table 3). Unfortunately, the higher reactor temperatures do not favor the Claus reaction. Consequently, the first catalytic reactor frequently is operated at a rather high temperature, while later reactors are operated at temperatures as low as possible. A typical three-reactor Claus unit will have the first reactor inlet at 450- 480°F, the second reactor at 390-430°F, and the third reactor at 370-410°F. This allows for maximum reaction of COS and CS2 in the first bed and maximum H2S-SO2 reaction in the second and third beds.

Typically, a temperature rise across each reactor occurs because the reaction is exothermic. The theoretical temperature increase generally ranges between 80 and 180°F for the first reactor, 20-60°F for the second reactor and 5-15°F for the third reactor. Actual measured plant temperatures are less than these values because of heat losses to the atmosphere.

MECHANICAL CONSIDERATIONS

Reaction Furnace. Design of the burner and furnace are very important since the operating temperature usually ranges between 1, 750°F and 2,600°F. These high temperatures require refractory lining to prevent overheating the steel vessel used in external furnaces. However, some furnaces place the main burner in the waste heat boiler where it is surrounded by water. This is called a "fire tube" unit and has no refractory. Normally the fire tube varies from a minimum of about 16 inches in diameter up to maximum of about 40 inches. The fire tube is the least expensive installation for small Claus units. However, the external reaction furnace is generally more practical for capacities over 30 LT/D.








Normal furnace residence time should be 0.5 to 1.5 seconds, depending on H2S concentration in the feed gas. Residence time less than 0.5 second can cause operating difficulties and inefficiencies.

Refractories capable of withstanding high temperature normally have relatively high alumina contents (> 60%). Unfortunately, such refractories, which are poor insulators, do not provide adequate protection for steel vessels. Consequently, a second refractory layer is frequently used to provide proper insulating qualities. These second refractories generally cannot withstand high flame temperature, thus forcing the two-layer system. The combination refractory is usually a minimum of 7 to 9 inches thick. Caution should be used in refractory design because overheated steel (> 650°F) results in direct sulfide attack while overcooled steel (< 300°F) causes acid corrosion in this system. Both phenomena cause premature failure of the furnace shell.


WASTE HEAT BOILER

Boilers other than large fire-tube type use smaller tubes which are exposed to hot combustion gases. Consequently, these tubes must be protected to ensure satisfactory life. Additionally, the inlet tube sheet is protected from direct contact with hot gases by use of about 3 inches of refractory.

The inlet end of each tube is shielded by inserting into the tube a ceramic ferrule that extends about 6 inches beyond the inlet tubesheet. This ferrule also protrudes about 2 to 3 inches in front of the inlet tubesheet to prevent hot gas from contacting the tube directly at the critical junction of tube and tubesheet. Use of ceramic ferrules is imperative if reasonable tube life is to be expected. For best results, these ceramic ferrules are wrapped with a thin layer of insulation (1/8-3/8 in. thick).

Tube size ranges from 2 to 6 inches for most boilers. Generally, welded (ERW) tubes are installed, but occasionally seamless tubes are used. Tube spacing is based on a minimum ligament dimension of 0.75 -1.0 in. Design mass velocity ranges between 1.0 and 8.0 Ib/sec/ft2 based on allowable tube-side pressure drop.

Tubesheet design (ASME code, Section I) for these vessels is greatly influenced by thermal/mechanical stress at the tube attachments. Tubesheets are usually thin (¾ to 1½ in.), resulting in acceptable temperature profiles at tube welds, tube sheet ligaments between tubes, and for portions not backed by water. Deflection that occurs during operation must be considered in order to reduce additional stress transmitted to the tube attachment. Boilers that are fully flooded have tube patterns arranged to account for bending stress at the outer boundaries of the tube sheets, whereas kettle-type boilers have stayrods to stiffen portions of tubesheets not supplied by tubes. Where possible, boilers should be provided with bolted inspection ports at the end and/or beginning of each pass.

Proper process design of the boiler requires that heat ex- change calculations include effects of radiation transfer as well as normal convection transfer. If radiation is ignored in calculating surface area requirements, the Claus boiler is usually appreciably over-sized. This can lead to operating problems because a waste heat boiler that is too large can cause premature sulfur condensation, improper reheat, viscous sulfur formation, etc.


SULFUR CONDENSERS

Sulfur condensers are built either as single-pass or multi- pass units (up to four tube passes in a single shell). In either case, they are generally designed with a minimum tube size of 1 inch. Normally 12 BWG steel tubes are used and installed with a ½ to ¾-in. ligament between the outer walls of the tubes. Welded (ERW) tubes are generally used.

Sulfur condensers operate at heat fluxes much lower than waste heat boilers and usually at much lower steam pressures. Both factors significantly reduce design criteria for tube sheets and tube attachments.

Sulfur condensers may be designed in accordance with ASME Section I or Section VIII Codes. In either case, tubesheet design need only meet ASME rules of design to pro- vide satisfactory service. Sulfur condenser tube attachments are typically rolled, seal welded, and re-rolled.

Design mass velocity varies from 3.0 to 8.0 lb/sec/ft2. Care must be taken to keep mass velocity sufficiently high during turndown conditions to prevent "sulfur misting." If "sulfur misting" occurs, liquid sulfur will not separate adequately from the process gas in the condenser separator chambers.

The liquid sulfur drain leg and nozzle should be steam- jacketed to prevent plugging by solid sulfur, and the drains should be installed for easy access for cleaning and disassembly. If possible, the sulfur condenser should be the low point in the Claus unit to allow any condensed sulfur to freely drain into ,the condenser and flow out through the drain legs.

Inspection ports and manways at each end of the condenser should be bolted to allow access to tubes for cleaning and inspection.

Vapor-liquid separator chambers are used to separate liquid sulfur from uncondensed process gas. These are made as either integral parts of the condenser or as separate vessels. Most sulfur condensers are designed to generate steam on the shell side while the gases are cooled and sulfur is condensed on the tube side. Design steam pressure is usually 20 to 80 psig.


CATALYST BEDS


Most catalytic reactors use a design flow of 2.0 to 4 mole& per hour of process gas per cubic foot of catalyst bed. Normally, activated alumina catalyst (density equals 45-53 lb/ft3) is supported at the bottom of the bed with a 3 to 6-inch depth of more dense material (84-100 lb/ft3). Support material prevents catalyst migration from the reactor to the sulfur condenser, can plug sulfur drain legs and condenser tubes, necessitating a plant shutdown for cleaning. Catalyst bed support material is usually provided with a carbon steel sup- port floor and stainless steel screen. A minimum of 3 inches of support material is also recommended on the top of the catalyst bed to serve as a gas flow distributor.

Frequently, catalyst beds are placed in a single horizontal vessel for units less than 100 LT/D. Each bed is separated by internal partition plates. For the larger units, individual catalyst beds are normally used. Although vertical reactors are used, they are not normally economical in units less than 800 LT/D. Internal refractory lining of these vessels is not necessary unless in-place catalyst regeneration is planned, in which case the support grating must be designed for elevated temperatures. Each catalyst bed should have a minimum of 3 inches of insulation outside if no internal refractory is used. When internal refractory lining is supplied, the outside insulation thickness can be reduced to 1 to 2 inches.

Nozzles located in the bottom of catalytic reactors should be installed flush with the vessel interior, otherwise internal projections act as weirs, allowing corrosive material to accumulate and corrode carbon steel walls.


PIPING

There are two major areas of piping concern: liquid sulfur drain lines and vapor process lines. Since sulfur freezes below 245°F, liquid sulfur drain lines must be adequately heated and insulated to prevent sulfur solidification. Main process lines are also heavily insulated to prevent vapor sulfur from condensing and/or subliming. However, the more critical concern in large process piping is allowance for thermal expansion.

Process lines expand because outlet temperatures from catalytic reactors range between 400 and 680°F. Expansion of these lines can cause severe damage to vessels, heat exchangers, and pipe if not adequately designed. Either stainless steel convoluted expansion joints or expansion "loops" are used on catalytic reactor outlets. Convoluted joints should be installed in vertical pipe sections only and should be insulated on the outside.

Carbon steel piping is adequate. Schedule 40 is normally used for small units. However, units requiring 14-inch or larger lines need only standard weight pipe. For process lines over 30 inches in diameter, piping thickness can be the same as for catalytic reactor vessel walls.

Typically, condensers are sealed by liquid sulfur, since it drains into the sulfur collection vessel. This prevents uncondensed gases from escaping to the atmosphere via the sulfur pit. Molten sulfur drain lines, which usually include both horizontal and vertical sections, need to have installed adequate facilities for clean-out to avoid premature unit shut-down for unplugging drain lines.

The best method of heating liquid sulfur drain lines is with jacketed pipe. Usually steam is used as a heat source, but hot oil is also satisfactory. For best performance, the system should be designed to maintain molten sulfur at a minimum temperature of 280°F. Generally, it is best not to use superheated steam because of the danger of overheating sulfur. Above about 325°F, the sulfur viscosity increases rapidly, preventing free flow from condensers to storage (Fig. 1).

Since liquid sulfur seals the condenser to prevent vapors from getting to the atmosphere, it is imperative that the "net" liquid column be sufficient to withstand the maxi- mum operating pressure, usually the air blower maximum discharge pressure.


INCINERATOR/STACK

All sulfur compounds can be oxidized to SO2 both thermally and catalytically. Thermal oxidation normally occurs at temperatures between 1,000 and 1,500°F in the presence of excess oxygen. The penalty for using insufficient oxygen and/or lower temperatures is that some H2S may not be converted to SO2; thus H2S is discharged to the atmosphere. Using too much oxygen consumes costly fuel.

Claus unit tail gas will not support combustion because there are too few combustibles present (<3%). Thus, oxidation of all sulfur components in tail gas to SO2 requires use of external fuel and air.

Catalytic incineration, a recent development, works when tail gas is heated to 750-900°F. Heated gas and heated air in contact with a fixed-bed catalytic reactor promote oxidation of sulfur components to SO2.

Air is fed by natural draft for most thermal incinerators, but for those operating with waste heat recovery sections, a blower is used. Usually at least 25% excess air is used for proper tail gas incineration.

Incinerator and stack are often combined into a single vessel, called an incinerator-stack. In this convenient design, the incinerator is usually an enlarged base and the stack of a small diameter extending upward from the incinerator. A horizontal burner normally mounts into the incinerator side. Tail gas can be fed either through the burner or into the incinerator adjacent to the burner via another nozzle.

Normally the incinerator is sized for at least 0.5-1.5 seconds residence time.

Most incinerator-stacks are constructed with a steel shell lined with heat-resistant refractory. Such stacks can be free- standing, or guy- or derrick-supported structures. Selection of support type is dictated by stack height, windload and/or, seismic criteria for the site. A guy-supported stack is most: economical; however, a larger plot space is needed for guy anchors and cables. A free standing stack is most commonly: used for heights not exceeding 250 ft. Taller stacks are usually derrick-supported.

Incinerators are generally designed for a maximum operating temperature of 2,000°F. Thus refractories rated as low as 2,200°F have been successfully used in incinerator-stack installations. Normally, one layer of refractory is sufficient, and is usually an insulating castable rated at service conditions of 2,200°F or above.

Vertical incinerators installed at the base of the stack require proper protection for the incinerator steel floor. The preferred protection is circulation of ambient air below the steel floor. For best results, the steel should be kept cooler than 650- 700°F, and hotter than 300°F. Above about 650°F, steel can be directly sulfided. Below 300°F, water can condense against the shell, absorb SO2 and form a very strong, highly corrosive acid. Accomplishing these temperature goals requires careful design of both inside refractory lining and outside insulation. Normally, refractory thickness to protect the incinerator shell (either vertical or horizontal type) is 2 to 4 inches. The incinerator floor on vertical units normally employs 4 to 6 inches of refractory.

The incinerator-stack should be insulated on the outside to prevent low shell temperature. This insulation can be an air gap (usually 3 to 4 inches deep) with either a stainless steel or aluminum sheathing around the incinerator-stack, or a <1.0 inch-thick layer of blanket-type insulation covered with aluminum or stainless steel. Over-insulation can cause steel shell overheating.

Most government air control units (federal, state and/or local) require that emissions be monitored on a periodic basis. Thus the incinerator-stack is usually equipped with plat- forms, sample nozzles and utilities located at the proper level. This equipment permits periodic measurement of stack flowrate and stack gas composition for use in determining sulfur emission rates.

Where continuous monitoring is required, SO2 stack gas analyzers are installed on the stack. These instruments are often coupled with flow meters to develop a continuous record of flow and SO2 content.

Typical stack velocities for design are 40 to 100 ft/sec, with allowable pressure drop generally determining actual design flowrate.