Indian Institute of Technology Roorkee

Department of Paper Technology

A report on:

Smelt Water

Explosions

Submitted to: Prepared by:

Dr. Ram Kumar Ashutosh Parihar

Balram Sharma

B.Tech IV

ACKNOWLEDGEMENT

We would like to thank Dr. Ram Kumar for giving us this opportunity to

write a report on Smelt Water Explosions - “History, Reasons, Mechanism and

Prevention”.

We would also like to thank all our classmates for providing lot of

information which are helpful for the report. This report on smelt water

explosions helped us in learning a lot about the types of explosions, there causes

etc.

CONTENTS

Introduction

History

Reasons

Mechanism

Prevention

Conclusion

Bibliography & References

INTRODUCTION

A smelt water reaction can occur anytime smelt and water come into

contact. However, not all instances of possible smelt water contact result in an

explosion of sufficient magnitude to damage the boiler or surroundings. One of

the critical factors is the location and magnitude of the water source.

The number of recovery boiler explosions has been steadily decreasing,

since 1960. Explosion can be caused by the combustible gas from either auxiliary

fuel or Pyrolysis gases or by smelt water interactions. The single biggest cause of

recovery boiler explosions is from a pressure part failure resulting in a smelt

water explosion.

Smelt water explosions result when leak allows water to build up on the

smelt bed. The extremely rapid build up of steam causes symmetric damage to

the lower furnace and asymmetric damage to the lower furnace. The mechanism

of smelt water explosions is not as well understood as for combustible gas

explosions.

The most serious smelt water explosions come from large leaks that allow

significant build up of water on the hearth. The exception to this rule is small

leaks in the floor tubes that introduce water directly into the smelt bed. Rapidly

draining the boiler of water and stopping water entry will prevent most

explosions once water has entered the furnace.

HISTORY

A break down of the recovery boiler explosion history is given in the figures.

In the 1980s, there were 20 recovery boiler explosions. During 1970s, there were

38 explosions. The number of explosions was nearly cut half in 1980s. This is a

real decrease in explosion frequency and not due to a decrease in the number of

boilers operating.

The data in the fig. 2 appear to suggest that black liquor explosions have

been decreasing and that Pyrolysis gas explosions have increased. This is

misleading. Black liquor can cause explosions in two different ways. The liquor

may contact the smelt in the bottom of the furnace and cause a smelt water

explosion. It may also lead to a furnace blackout, formation of Pyrolysis gases

from black liquor, and a combustible gas explosion. During 1960s and 1970s, black

liquor Pyrolysis gas explosions were not recognized as a cause for recovery boiler

explosions, and all explosions connected with the liquor system were classified as

smelt water explosion due to weak black liquor. It is likely that many of these

weal liquor explosions were actually Pyrolysis gas explosions. There probability is

little real difference in the Pyrolysis gas explosion experience in the 1970s and the

1980s.

The

frequency of auxiliary fuel explosions has dropped by nearly an order of

magnitude, despite the large expansion in Kraft pulp capacity that has occurred

since 1965. Auxiliary fuel explosions accounted for 45% of all the recovery boiler

explosions up through 1965. Since then, only 12% of all the explosions have

involved auxiliary fuel. The substantial drop in auxiliary fuel explosions from 1970s

to the 1980s is a continuation of a trend that began in the 1960s with the

introduction of monitored burners for recovery boiler service.

Most

recovery boiler explosions are smelt water explosions. Pressure part failures are

the biggest single cause of recovery boiler explosions in both the 1970s and the

1980s, but the number of damaging explosions from this cause was reduced by

about half in the 1980s.

REASONS

Explosions became a greater problem as the Kraft industry grew after

World War II. There was considerable controversy about the nature of smelt

explosions. Many people thought smelt water explosions were combustible

explosions involving hydrogen and possibly sodium metal. That idea is now

thoroughly discredited.

The first published study of the smelt water system was an investigation of

dissolving tank explosions that found pure Na2CO3 to be non explosive and that

the addition of NaCl and NaOH made the smelt explosive. In more than 500 smelt

quench experiment, no flames or gaseous combustion were observed.

In the early 1960s, a large number of experiments on smelt water

explosions were carried out by Babcock & Wilcox and Combustion Engineering

under sponsorship of the Smelt Research Group. This work was summarized as

follows. “Research has identified the smelt water reaction as an extremely rapid,

extremely intense physical explosion consisting of the explosively rapid

generation of steam. However, the research has not been encouraging with

respect to the possibility of rendering molten Kraft smelts non explosive with

water solutions, by means compatible with the chemistry and economics of the

Kraft process. Thus as a practical matter, our principal hope for reducing the

incidence of smelt water explosions in the smelting recovery furnaces lies in

keeping molten smelt and water phases apart.” This situation is still true today.

This research on smelt water explosion reconfirms the effects of smelt

composition on explosivity found in the earlier work. Sodium Carbonate was not

explosive, while sodium chloride, sodium hydroxide, and sodium sulfide would

make smelt explosive when added to sodium carbonate. Cold water reacted more

violently than hot water and green liquor reacted more violently than plain water.

However, work in Sweden in which blasting caps were used as detonators to

trigger the explosions showed that an explosive interaction between sodium

carbonate and water could be obtained.

Research on smelt water explosions proved that explosions were physical

steam explosions, but did not result in any good theoretical understanding of how

the explosive steam generation occurred. The energy needed to make steam is

present in large excess in the bed and smelt, but how does this energy make a

large amount of steam in a very short time. An explosion requires coordinated

extremely rapid vaporization of a large amount of water. There are two ways this

can occur:

1. The energy transfer from smelt to water occurs relatively slowly (compared

to the time scale of the explosion) and is stored in an unstable state from

which it can be suddenly released.

2. The energy transfer from the smelt to the water occurs at the same time as

the blast is developing and is closely coupled to it.

MECHANISM

Smelt water explosions generally are steam explosions. The mechanism

which is greatly accepted world wide, explains that explosions consists of 4

stages:

1. Coarse Premixing:

The hot and cold fluids become intermixed on a scale small enough to

permit coupling with the blast wave. The configuration is maintained in a

quasi stable state by film boiling. The intermixing occurs on a time scale

that is slow compared to the time of the explosive interaction.

2. Triggering:

An event collapses the vapor separating the hot and cold fluids and initiates

direct liquid contact in a localized region. The pressure pulse developed

acts as the initiator of the blast wave.

3. Escalation:

The localized interaction escalates into a full fledged detonation wave. The

pressure pulse propagating through the quasi stable mixture of hot and

cold fluid collapses the stabilizing vapor film and causes fine fragmentation

and rapid heat transfer. This generates rapid vaporization that feeds

pressure energy into the blast wave.

4. Expansion:

The blast wave moves out and does mechanical work on the surroundings.

According to this theory, if the smelt and water are in close proximity to

each other and a small scale explosion is initiated, the resulting shock wave will

cause the fragmentation and rapid heat transfer needed to give a large scale

explosion.

The spontaneous nucleation concept and the 4stage theory of steam

explosion have been tied together in one theory. A liquid-liquid contact

temperature above the spontaneous nucleation temperature is necessary for

establishing film boiling which allows the coarse intermixing to proceed to the

point where a coherent explosion is possible. The contact temperature is usually

estimated by using the solution for the time independent contact temperature

between two slabs of materials each initially at a uniform temperature.

PREVENTION

Prevention of smelt water explosions is dependent on preventing failures that

allow water to enter the furnace and by following the procedures that minimizes

the amount of water entering the furnace and the length of time that this occurs

when a tube leak emergency develops.

1. Emergency Shut down Procedures

BLRBAC has developed a recommended emergency shutdown procedure.

An immediate emergency shutdown must be performed whenever water in

any amount is known or suspected to be entering the furnace and cannot

be stopped immediately. The essential elements of the emergency

shutdown procedure are the following:

a) Activate audible and visible alarms to clear the recovery boiler area of all

personnel.

b) Stop firing all fuel. Divert black liquor. Secure auxiliary fuel at a remote

location.

c) Shut off the air supply to the primary air ports. Provide a balanced draft

and an air supply above the char bed to purge gases from the furnace.

d) Shut off feed water to the boiler.

e) Drain the water walls to a level eight feet above the low point of the

furnace floor.

f) Reduce the steam pressure as rapidly as possible as soon as boiler has

been drained.

The most critical function is alarming and clearing the area of all personnel. By

keeping the people away from the vicinity of the boiler while an emergency

condition exists, the risk of serious injuries or fatalities is minimized.

The boiler is rapid drained over a period of about 20 minutes to stop water

entry to the furnace. Shutting off feed water to the boiler contributes to the

same end. In order to prevent tube overheating when the boiler is drained, it is

necessary to immediately stop firing all fuel. A balanced draft and some air

flow are maintained to provide a continual purge of the furnace and help

prevent a buildup of combustibles in the furnace. Air flow to char bed must be

stopped, because continued bed burning and smelting could result in smelt

contacting and damaging of floor tubes.

The rapid drain step is unique to recovery boiler emergency shutdowns.

Experience shows that it has been effective in reducing the probability and

violence of smelt water explosions. Although the intent of a rapid drain is to

prevent explosions, it provides additional benefits as well. As the boiler is

drained and depressurized, the potential for violent release of the energy

represented in the high pressure steam and water is diminished and is

essentially non existent when the ESP has been completed. Even if a recovery

boiler explosion occurs and the boiler pressure parts are opened up, the

energy release outside the boiler will be minimized.

2. Leak Prevention

The most effective way to avoid smelt water explosions from tube leaks is to

prevent the tube failures from occurring in the first place. The record shows

that most dangerous types of tube failures are the big leaks that result in large

amounts of water entering the furnace and small leaks in the floor or in the

wall close to the hearth. The most common cause of tube ruptures is short

term overheating caused by lack of cooling water in the affected area. Many of

these over heat failures have been caused by operating the boiler for an

extended period with no feed water or with limited feed water and the

restoring water flow. Such failures are effectively prevented by low drum trips.

Widespread use of low drum trips has greatly reduced the frequency of

overheats failures. Pluggage of tubes with debris, particularly after the repair

work, has also led to a significant number of overheat failure. General wastage

from corrosion can also lead to a tube failure. Most mills have effective

programs in place for pressure part inspection and tube thickness monitoring

so that action can be taken before tube thin to the point where tube failure is

likely. Improved water quality has reduced both short term overheat failures

from tube Pluggage by water side sludge and also corrosion caused by high

metal temperatures when deposits are present. Many of the smaller leaks

have been caused by weld problems, particularly at welded attachments.

Improvements in welding procedures and in welding quality control have

helped reduce the frequency of such failures.

One of the common causes of the tube ruptures is rapid thinning from

accelerated corrosion due to being sprayed with water from a small leak on an

adjacent tube. This thinning occurs over a period of hours, so early detection

of small leak can prevent a much larger tube rupture.

3. Leak Detection

Early leak detection is pivotal in recognizing a smelt water emergency and

activating the alarm system to evacuate people from the danger area. A strong

case can be made for having the ability to activate alarms independently of the

rest of the Emergency Shutdown Procedure. If there is uncertainty as to

whether or not an emergency situation exists, the alarms can be activated and

the area around the boiler evacuated while the situation is being analyzed.

Tube leak emergencies tend to come in two broad classes; large failures

that results in water input rates of hundreds of gallons per minutes, and small

leaks that put in only a few gallons per minute. Leak detection requirements

are different for these two types.

If a large tube failure occurs and the boiler does not trip out, the leak is

relatively easy to recognize. There will be an increase in feed water flow, a

drop in drum level, a large increase in the feed water flow/steam flow

differential, an increase in furnace pressure (positive draft), and an increase in

ID fan speed (if not already at the maximum). There may also be loud roaring

noise heard and steam, smoke, and fire may be coming out of the furnace

openings.

If a tube rupture causes the boiler to trip out, either form high furnace

pressure or low drum level, the indicators are more complex. Whenever there

is a loss of fire in the furnace the saturated steam water mixture in the boiler

drum shrinks due to the increase in the volume of bubbles. This shrinkage will

cause a falling drum level and increase in the feed water flow/steam flow

differential to differentiate a tube leak from a normal trip. The loss of fire will

also decrease the load on the ID fan and so the furnace pressure may recover

quickly and the boiler may not blow back for any length of time. In this case, it

may take a longer time to recognize that water is entering the furnace and

that emergency action is needed.

Smaller leaks are harder to detect. Feed water flow/steam flow differential

is good indicator if it can be picked up over the noise in the signals,. Local black

areas in the bed or an unexplained rise in TRS may also be indicators of water

reaching the char bed.

Technology is now available to help operators detect leaks. For boilers that

have distributed control systems, the average feed water flow and steam

differential can be sensitive way to pick up leak indications and provide an

alarm. There is a relation between the size of leak that can be detected in this

manner and the length of time used for averaging. Longer time periods are

needed to detect smaller leaks. A case can be made for using both a short time

period to provide sensitivity for small leaks. There are two systems in use to

detect leaks. One uses statistical analysis of the feed water flow and steam

flow data. The other uses an expert system approach.

Acoustic leak detectors have been used on recovery boilers for over ten

years with some success. Two types of acoustic leak detector systems are in

use. One type has a sensor directly open to the furnace atmosphere and

listens for airborne sound in the furnace. The other type has acoustic sensor

attached to one end of metal rod which is attached to one end of metal rod

which is attached to the outside of the boiler wall membrane. This type

detects metal borne sounds of a leak. Both types have been able to detect

leaks that occurred within a certain distance of the probe. The metal borne

systems are easier to maintain. Air borne systems have not been very effective

in the lower furnace because of the very dirty atmosphere that exists and the

large amount of extraneous noise is present.

Another way to detect leaks is through boiler water chemistry monitoring.

A tube leak in the system generating circuit acts like an increase in blowdown

flow and causes a drop in boiler water chemical residuals. The rate of drop in

chemical concentrations depends on the leak rate relative to the volume of

water in the boiler. Thus, there is a relationship between the size of a leak and

the time needed to get a large enough drop in concentration to provide

reliable detection. This time depends on the size of the boiler.

4. Safe Firing of Black Liquor

Smelt water explosions have also involved water or weal black liquor

that enters the furnace through the black liquor system. BLRBAC has

developed procedures for safe firing of black liquor to prevent such incidents

from occurring.

The heart of the black liquor safe firing system is continuous monitoring of

the black liquor solids being supplied to the furnace and automatic diversion of

the black liquor from the furnace header if the liquor solids go below some

specified minimum value (currently 58% solids). The procedures includes

startup and tripping logic for black liquor firing analogous to that used for

auxiliary fuel. Since it is necessary to override the diversion system when

washing black liquor out of the header when firing has been discontinued, the

safe firing system also has interlocks that do not open unless the liquor guns

are proven to be physically removed from the furnace. The procedure does not

allow permanent water connections to the liquor system downstream of the

final solids monitor and recommends the use of removable spool pieces for

wash water connections.

The minimum safe firing concentration is sometimes mistakenly interpreted

as the solids content above which black liquor cannot react explosively with

smelt. The likelihood of an explosion and the violence of an explosion and the

violence of an explosion decreases as the solid content rises, but there is no

value that should be considered absolutely safe. There has been at least one

explosion involving liquor with a solids content of 68%. This occurred when a

continuous steam of liquor roped into the smelt pool during upset operation.

Other variables than dry solids content are also important in black liquor smelt

explosions. These include viscosity, the amount of liquor, the smelt liquor

contact geometry, and the possibility of disturbance acting as triggers.

The procedure for safe firing of black liquor also includes recommendations

for smelt spout cooling water system design, operation, and maintenance. The

recommended design is for a closed loop system with treated water. Either a

low pressure gravity flow or a vacuum (siphon) flow system can be used.

Cooling water flow and temperature are to be monitored and alarmed.

Systems should not allow shutting off water on the downstream side of the

spouts.

Smelt spout openings may plug occasionally due to low loads of other

operating conditions. This is a potentially dangerous condition. If smelt is

allowed to accumulate in the furnace behind the plugged opening, heavy smelt

flow may occur when the opening is unplugged and result in a dissolving tank

explosion.

The BLRBAC safe firing of black liquor procedure also contains

recommendations to prevent dissolving tank explosions. The BLRBAC

recommendation include a fixed, high level suction point for the grenn liquor

pumps, adequate steam shatter jets for all spouts with a backup system for

emergency shattering, adequate agitation (with backup) to green liquor

density control system, and provisions for emergency dilution water that is

available in a power failure.

CONCLUSION

The smelt water explosion experience in recovery boilers in

consistent with the current theory of so called steam explosions. The key concept

in this theory is that smelt and water must accumulate in such a way that

relatively large amounts can be suddenly brought into intimate contact and a

shock wave initiated. In this manner, very large amounts of steam can be

generated in a few milliseconds as the shock wave propagates through the

mixture. This feeds a great deal of energy into the developing shock wave and

result in a violent explosion. The presence of a large amount of hot char and

frozen smelt may add to the energy release when the explosion is initiated.

The explosion experience is consistent with the idea that some pudding of

water on the hearth is necessary for a major smelt water explosion. The large

pressure part failures where the water can ether the furnace cavity and small

leaks in the floor are the types of failures where accumulation of water on the

hearth is most likely.

There are some reasons why immediate explosions from pressure part

failures are unlikely. The water entering the furnace cavity will flash. The amount

of flashing depends on the boiler pressure. It is about 30% at 600 psi, 37% at 1000

psi and 45% at 1500 psi. This is the amount of water that will flash to steam. Part

of the remainder will evaporate by hot gases in the furnace. A pressure part

failure that directs the steam/water spray to a wall where the water can run

down to the hearth, or a large failure high in the boiler directed down at the bed

is probably much more dangerous.

BIBLIOGRAPHY & REFERENCES

1. www.google.com

2. www.about.com

3. www.paperonweb.com

4. www.tappi.org