Search

Mud ball formation in lime kilns has been a persistent problem in many kraft mills, particularly for older kilns that are equipped with chains. A systematic laboratory study was conducted to examine how mud balls are formed and the key factors that affect ball formation. The results confirm the general mill experience that high moisture and high sodium contents in lime mud are the main contributing factors to ball formation. The high moisture content allows lime mud to agglomerate and grow to form balls, while the high sodium content helps make the balls hard and retain their shape. A ball formation mechanism is proposed to explain how mud balls form and grow near the kiln feed end.

A study was conducted to examine the most likely parameters responsible for poor dregs settling at a kraft mill over a 2.5-year period, using multivariate data analysis (MVDA) and machine learning (ML) techniques. The dregs settling behavior seems to be seasonally influenced, implying that wood quality variation can be a factor. The results from the MVDA/ML analysis show that poor dregs settling is correlated to incomplete combustion and/or low load conditions in the recovery boiler, low sulfidity in the causticizing plant, and high flow in the green liquor–weak wash cycle. Compositions of dregs and black liquor were also examined to identify correlations with impaired dregs settling. The results show that poor dregs settling strongly correlates with high silicon (Si) content in dregs and moderately correlates with high iron (Fe) and high aluminum (Al) contents, and with low bulk density in dregs. For mills that experience dregs settling or green liquor filtering issues, regular compositional analyses of dregs, green liquor, weak wash, and black liquor are recommended in order to monitor the dynamics of silicon and other constituents in the recovery cycle.

The lime kiln at the Arauco Constitución mill experienced severe ringing problems requiring it to be shut down for ring removal every 3 to 6 months. The mill controlled the problems by blasting ring deposits off during operation with its existing industrial shotgun and a newly installed Cardox liquid carbon dioxide (CO2) cartridge system. Various ring blasting procedures were tested to determine the optimum ring location and thickness to blast; the optimum depth to insert the CO2 cartridge into the kiln; and the most effective blasting frequency and sequence to employ. The best strategy was found to be the weekly blasting operation that alternated between the liquid CO2 cartridge and the industrial shotgun, with the CO2 cartridge inserted into the ring mass, 20 cm (8 in.) away from the refractory brick surface, and the shotgun aimed at rings at about 28 m (92 ft) from the kiln discharge end. With each blasting event removing considerably more rings than before, it takes a longer time for rings to rebuild, allowing the kiln to run continuously between annual maintenance shutdowns with only a few short (< 4 h) downtimes for ring removal. This substantially reduces the costs associated with ring removal and lime replacement during unscheduled shutdowns.

The effective shattering of molten smelt is highly desired in recovery boiler systems. Ideally, shatter jet nozzle designs should: i) generate high shattering energy; ii) create a wide coverage; and iii) minimize steam consumption. This study proposes a novel converging-diverging multi-jet nozzle design to achieve these goals. A laboratory setup was established, and the nozzle performance was evaluated by generating jet pressure profiles from the measurement of a pitot tube array. The results show that the shatter jet strength is greater with a large throat diameter, high inlet pressure, and a short distance between the nozzle exit and impingement position. Increasing the number of orifices generates a wider jet coverage, and the distance between the orifices should be limited to avoid the formation of a low-pressure region between the orifices. The study also demonstrates that an optimized converging-diverging multi-jet nozzle significantly outperformed a conventional shatter jet nozzle by achieving higher energy and wider coverage while consuming less steam.

The formation of hard calcite (CaCO3) scale in green liquor handling systems is a persistent problem in many kraft pulp mills. CaCO3 precipitates when its concentration in the green liquor exceeds its solubility. While the solubility of CaCO3 in water is well known, it is not so in the highly alkaline green liquor environment. A systematic study was conducted to determine the solubility of CaCO3 in green liquor as a function of temperature, total titratable alkali (TTA), causticity, and sulfidity. The results show that the solubility increases with increased temperature, increased TTA, decreased causticity, and decreased sulfidity. The new solubility data was incorporated into OLI (a thermodynamic simulation program for aqueous salt systems) to generate a series of CaCO3 solubility curves for various green liquor conditions. The results help explain how calcite scale forms in green liquor handling systems.

Pulp and paper mill biosludge becomes sticky after being dried to a certain solids content. As biosludge becomes sticky, it agglomerates and adheres to the heat transfer surfaces of the dryer. This undesirable property can lower the dryer efficiency and cause the drying equipment to fail. A systematic study was conducted to examine the sticky behavior of biosludge. The stickiness was evaluated by measuring the adhesive force between a sludge cake and a stainless steel substrate, and the cohesive force between a sludge cake and a sludge substrate. The results show that: i) both adhesive and cohesive forces increase markedly as the solids content increases, reaching a maximum value at about 13% solids, and then decrease steadily at a higher solids content; ii) cohesive force is stronger than adhesive force, implying that biosludge tends to agglomerate rather than adhere to smooth equipment surfaces; and iii) mixing wood fines or fly ash from a biomass boiler reduces the stickiness of the mixture. These findings may help mills improve the thermal efficiency of biosludge dryers and to turn biosludge into a more attractive fuel for burning in biomass boilers.

The explosion energy generated as molten smelt droplets interact with water was evaluated as a function of smelt distribution, water temperature, and smelt temperature using a thermodynamic model. The results show that increasing smelt-to-water volume ratio and water temperature significantly increases the explosion energy, converting a larger proportion of the thermal energy of smelt into mechanical work. To reduce the chance of violent smelt-water explosions, it is important to: i) optimize the shatter jet design and operation to uniformly distribute the smelt over a large area in the dissolving tank; ii) avoid high green liquor temperature and ensure adequate liquor mixing; and iii) avoid upsets that may cause heavy smelt runoff or jellyroll smelt.

Electrostatic precipitators (ESPs) are used in most pulp mills to remove particulate from recovery boilers, power boilers, and lime kilns. As environmental regulations have become increasingly stringent in recent years, maintaining high ESP performance is of vital importance in mill operation. This paper discusses results of a literature review of the ESP technology used in industrial combustion units, including recovery boilers, as well as results of a parametric study using the well-known Deutsch-Anderson equation to correlate recovery boiler operating conditions with ESP collection efficiency. The results show that for particles up to about 0.3 μm, the ESP collection efficiency decreases drastically with increased particle size and with decreased temperature. For particles larger than 0.5 μm, however, the trend reverses; the collection efficiency increases with increased particle size and decreased temperature. The results also suggest that the particle concentration (or loading) in the flue gas has no effect on collection efficiency and that sodium chloride particles are more readily captured than sodium sulfate particles. The latter prediction, however, appears to be in contradiction with mill experience that sodium chloride particles are more difficult to capture.

Ring formation in lime kilns is a common problem in pulp mills. Unstable burner operation that causes wide temperature variations within the kiln has been shown to be a main contributor. As rings grow thicker and longer, they are expected to affect the burner flame pattern, alter the temperature distribution, and further aggravate the problem. This study systematically examines the effect of rings on burner operation as a function of primary air and ring characteristics, using a laboratory mock kiln. The results show that primary air greatly affects the burner flame pattern and stability. Without a ring, the flame is brighter and shorter, with an increase in primary air (PA) up to 17%, and blows out at 20% PA or higher. In the presence of a ring, the flame becomes unstable and blows out when the ring is placed near the burner, but it is more stable when the ring is placed away from the burner. Thick and long rings destabilize the flame more than thin and short rings.

Handling and disposal of biosludge has been a persistent problem for many pulp and paper mills primarily because of the material’s high moisture content and poor dewatering and drying characteristics. Most biosludge is landfilled or incinerated, and some is used as a soil amendment. To incinerate biosludge, the mills mix it with primary sludge, mechanically dewater the mixture to 15%–30% solids, and burn it with hog fuel in biomass boilers. For biosludge to burn effectively, however, it must be dewatered or dried to at least 30%–35% solids. A systematic study was conducted to examine the drying characteristics of biosludge collected from pulp and paper mills at various temperatures using a thermogravimetric dryer. The drying behavior of the tested biosludges was similar, adding fillers such as wood fines and biomass boiler fly ash to biosludge up to 30 wt% did not significantly affect the drying rate, higher organic content might make biosludge more difficult to dry, and increasing drying temperature significantly increases the drying rate as expected.

Operating data of a bubbling fluidized bed (BFB) boiler and three stoker grate (SG) biomass boilers from different pulp mills were analyzed over a 2-year period. The results show that in all cases, the thermal perfor - mance decreases markedly from 5.5 to 4 lb steam/lb dry biomass as the feedstock moisture content increases from 40% to 60%. The BFB boiler had better thermal performance, although it operated in a higher moisture content range compared with the SG boilers. Multivariate analysis was also performed on one of the SG boilers to determine operating parameters that affect thermal performance. The results show that furnace temperature, oil flow rate, and induced draft fan current positively correlate with thermal performance, while the feedstock moisture content, total air flow, and excess oxygen (O 2 ) negatively correlate with thermal performance. This implies that when making mod - ifications to improve thermal performance, it is important to take into account correlations among various parame - ters. In some cases, one positively correlated parameter might cause an increase in a negatively correlated parame - ter. The net effect might be a decrease in thermal performance.

Molten smelt normally flows smoothly down the smelt spout of a recovery boiler like water, but at times it suddenly becomes sluggish and forms a viscous blob on the spout trough that partially or completely blocks the smelt flow. This form of smelt is commonly referred to as “jellyroll” smelt. How such smelt forms has been a puzzle to boiler operators and mill personnel for years. Numerous mill observations and the results of a recent study performed on both smoothly flowing smelt and jellyroll smelt collected from a recovery boiler suggest that that jellyroll smelt can form through three main mechanisms: i) the freezing of the molten smelt, ii) the melting of fallen deposits, and iii) the inclusion of a large amount of unburned char in the molten smelt. These mechanisms are consistent with mill experience that jellyroll smelt tends to form in older recovery boilers burning liquor with low solids and low sulfidity.

Black liquor is often mixed with various types of materials before being burned in a recovery boiler to meet specific needs of kraft pulp mills. A systematic study was conducted using a thermogravimetric combustor to examine how added materials might affect the combustion behavior of black liquors obtained from several pulp mills. The results show that adding soap, caustic, white liquor, and sawdust significantly reduces the liquor swelling tendency, thereby requiring a longer time for the liquor to burn completely. Adding makeup saltcake, precipitator ash, sodium sulfate, and biosludge, on the other hand, has little or no effect on the liquor combustion behavior.

Fireside deposits in recovery boilers are typically white, red, pink, grey, black, or occasionally yellow, depending on where they are in the boiler, the mechanisms by which they are formed, and the environment to which they are exposed. Although rare, blue deposits have been reported, and some were “bluer” than others. This study systematically examines the cause of the blue coloration of deposits in recovery boilers. The results show that for a deposit to become blue, it must a) contain sodium carbonate, b) contain a small amount of manganese, c) be molten or partially molten, and d) have exposure to an oxidizing atmosphere. Because deposits always contain sodium carbonate and manganese, these requirements suggest that blue deposits can form only in the superheater region of the recovery boiler when oxidizing conditions prevail. Blue coloration is thus more likely to be observed in boilers operating at a reduced firing load with a high excess oxygen target.

Burning feedstock with high and constantly varying moisture content is a challenge in the operationof the stoker-grate type biomass boilers in pulp and paper mills. A fundamental study was performed, using a thermogravimetric combustor controlled at different temperatures, to systematically investigate the effect of moisture content on the combustion behavior of several woody biomasses. The results show that combustion occurred through three stages: drying, devolatilization, and char burning. Increasing the moisture content of the sample above 30% significantly increased the drying time and the time to ignition of the sample. Moisture content, however, had only a small effect on the devolatilization rate and virtually no effect on the char burning rate. Furnace temperature had a great effect on combustion. Samples containing 40% moisture could still ignite and burn readily at 800°C, but they took a much longer time to ignite at 500°C and did not ignite at 400°C. These results imply that in biomass boiler operation, the high moisture content in feedstock delays combustion and causes the furnace temperature to decrease. The low temperature, in turn, suppresses combustion, further decreasing the furnace temperature and causing more delay in combustion. This downward spiral will eventually lead to a boiler blackout unless the biomass feed rate is significantly reduced or the auxiliary fuel (natural gas and fuel oil) flow rate cofired in the boiler is increased to keep the furnace temperature high.

Overliming and the consequent presence of unreacted lime (or free lime) in the lime mud are commonly believed to be the cause of many problems in the operation of causticizing plants and lime kilns in kraft pulp mills. The free lime content in lime mud is typically determined in mill laboratories using a so-called ammonium chloride (NH4Cl) method and in commercial laboratories using a thermal decomposition (TD) method. Over the years, we analyzed many lime mud samples from mills and found that the free lime content was consistently low, < 3 wt% calcium hydroxide (Ca(OH)2), even in cases where overliming was suspected to have caused problems. A systematic study was therefore conducted to investigate the validity of free-lime measurement methods, the reason for the consistently low free lime content in lime mud, and if free lime values can be used to indicate overliming. The results show that the NH4Cl method is not suitable for determining free lime. The TD method is good, but the possible interference of magnesium hydroxide (Mg(OH)2) must be taken into account. Since most pulp mills perform their free lime analysis on mud samples collected from pre-coat filters which have been washed, the resulting free lime value is low, and thus, cannot be used to assess the extent of overliming in the causticizing plant.

Lime kilns are used to calcine lime mud to produce lime for reuse in the causticizing plant of pulp mills. With significant increase in fossil fuel costs and the sustained drive to minimize carbon footprint, there has been a great deal of research and development of the biorefinery concept and production of biofuels for use within pulp mills. Several new biofuels will become available, together with traditional mill derived fuels such as tall oil. In considering use of biofuels for replacement of fossil fuels in the lime kiln, each fuel has a different impact on kiln performance. An assessment on kiln performance can be achieved and compared to either heavy fuel oil or natural gas. In most cases, the lime kiln fuels with high oxygen content and moderately low calorific values burn in a similar manner as natural gas; hence, in assessing impact on kiln performance, production and induced draft fan limits together with kiln exit gas temperatures need to be considered.

For kraft pulp mills that have thermally limited recovery boilers, lignin removal from black liquor has become an attractive option for increasing pulp production by allowing more black liquor to be processed through the boiler. This study systematically examined the combustion characteristics of lignin-lean black liquor and precipitated lignin from three kraft mills using a thermogravimetric combustor. The results confirm that adding lignin-lean black liquor to its original black liquor decreased the heating value and the degree of swelling of the mixed liquor. The effect on liquor swelling, however, was insignificant for mixed liquors that contained less than 20 wt% of lignin-lean liquor. As with other biofuels, the combustion of precipitated lignin was found to occur through three main stages: drying, volatile burning, and char burning. During the volatile burning stage, hardwood lignin swelled significantly, softwood lignin did not swell much, and mixed hardwood and softwood lignin was somewhere in between. Although the char content in lignin was about half of the volatile content, it took 10 times longer for the char to burn compared to the volatiles.

There is considerable interest in the nature and causes of heavy smelt runoff from recovery boilers because of the role it has played in numerous dissolving tank explosions and for personnel safety around the dissolving tank. Most mills have experienced runoff problems, which can be caused by cleaning plugged spouts, burning down a large bed, low sulfidity smelt, startup with a bed in the unit, and improper firing practices. The peak smelt flow during a runoff is often 3 to 5 times normal and may be much greater in severe cases. Heavy runoffs are self-limiting and typically last less than 30 min. The geometry of the lower furnace plays a significant role in runoff events. Sloped floor units are more vulnerable to smelt pool buildup and heavy runoff when released. Decanting bottom units are inherently more tolerant of smelt pool buildup. Low sulfidity results in smelt with a high melting temperature, making the smelt easy to freeze and difficult to flow. Sulfate-rich slag/deposits falling on the hearth from the upper furnace can contribute significantly to runoff problems by causing dams and spout plugging, increasing the load of smelt pool to be removed, decreasing bed temperatures, and lowering the sulfidity of the smelt leaving the furnace.

Sootblowing technology used in recovery boilers originated from that used in coal-fired boilers. It started with manual cleaning with hand lancing and hand blowing, and evolved slowly into online sootblowing using retractable sootblowers. Since 1991, intensive research and development has focused on sootblowing jet fundamentals and deposit removal in recovery boilers. The results have provided much insight into sootblower jet hydrodynamics, how a sootblower jet interacts with tubes and deposits, and factors influencing its deposit removal efficiency, and have led to two important innovations: fully-expanded sootblower nozzles that are used in virtually all recovery boilers today, and the low pressure sootblowing technology that has been implemented in several new recovery boilers. The availability of powerful computing systems, superfast microprocessors and data acquisition systems, and versatile computational fluid dynamics (CFD) modeling capability in the past two decades has also contributed greatly to the advancement of sootblowing technology. High quality infrared inspection cameras have enabled mills to inspect the deposit buildup conditions in the boiler during operation, and helped identify problems with sootblower lance swinging and superheater platens and boiler bank tube vibrations. As the recovery boiler firing capacity and steam parameters have increased markedly in recent years, sootblowers have become larger and longer, and this can present a challenge in terms of both sootblower design and operation.

The electrical resistivity of ash particles is an important parameter that determines the efficiency of electrostatic precipitators. This systematic study examines the resistivity of recovery boiler precipitator ash as a function of electrical field strength, time of exposure, particle composition, and gas composition and temperature. Synthetic ash and actual ash samples from several pulp mills are used. The results show that most ash samples tested had a resistivity between 109 and 1010 Ω·cm, but one of the samples had an unusually high resistivity, 1012 Ω·cm. The resistivity increases with temperature up to about 140°C, then decreases. At a given temperature, the resistivity decreases with increasing moisture and sulfur dioxide concentration in the gas. Resistivity also increases with an increase in chloride content in the ash, but is not affected by the carbonate, sulfate, and potassium contents. The results imply that recovery boilers burning liquors with high solids and high chloride contents produce ash with higher resistivity, making it more difficult for electrostatic precipitators to capture.

A systematic study was performed in the laboratory and in the field to examine the effect of magnesium oxide additive on deposit accumulation and removal. Laboratory results show that the additive has little effect on the amount of deposits collected on a probe, but it can make deposits easy to remove if a coating layer thicker than 30 μm can be effectively applied on the probe surface. The results also show that mixing the additive with black liquor or injecting it separately has no significant effect on deposit accumulation and removal. Tests performed in a recovery boiler where the additive was continuously injected show no evidence for the existence of the coating layer. The magnesium oxide content in the boiler deposits varied between 0.1 and 0.4 wt%, which is several times lower than the amount that was found to be effective in laboratory tests. Nonetheless, the additive was shown to be effective in facilitating deposit removal if it can form a coating layer on deposit/tube surfaces. This may be possible by periodically injecting a large amount of additive into the boiler for a short period and reducing the black liquor firing rate at the same time.

The effects of flashing on black liquor spray patterns were investigated by analyzing numerous spray images obtained from laboratory experiments using small scale splashplate nozzles with water and experiments using actual size splashplate nozzles with black liquor. The results showed that flashing produces small droplets and increases droplet velocity. The liquor mass flow rate varies with direction: the rate is higher at the center than at the sides of the spray sheet, particularly at a lower excess temperature. At a higher excess temperature, however, the mass distribution becomes more uniform across the spray sheet. Criteria were developed for predicting the onset of flashing and for estimating the mean droplet size of the black liquor spray under flashing conditions.

Scaling, the formation of hard deposits, in green liquor handling systems is a persistent problem in many kraft pulp mills. Scaling is commonly believed to be a result of pirssonite (Na2CO3·CaCO3·2H2O) formation when the green liquor total titrateable alkali (TTA) is high. In this study, we characterized 12 scale samples obtained from 10 kraft pulp mills using various analytical methods, including TGA/DSC (thermogravimetric analysis/differential scanning calorimetry), XRF (X-ray fluorescence spectrometry), XRD (X-ray diffraction spectrometry), SEM (scanning electron microscopy), and EMPA (electron microprobe analyzer). The analysis identified only four of these samples as pirssonite; the remaining consisted mostly of calcite (CaCO3). The reason for the predominant presence of calcite in the majority of the scale samples is not known. It may be a result of selective dissolution of sodium carbonate (Na2CO3) from the pirssonite scale during the time when the green liquor total TTA was low, leaving the insoluble CaCO3 behind.

Two additional field trial studies were conducted in different kraft recovery boilers in Sweden using force-measurement probes to determine the force of sootblower jets under various blowing conditions. The results confirm the findings of previous trials that, at a given distance, the force exerted on a target by a fully expanded sootblower jet increases nearly linearly with an increase in lance pressure. At a given lance pressure, the jet force decreases drastically as the distance between the sootblower nozzle and the target increases. At a distance farther than 1 m from the nozzle, the jet retains less than 10% of its original strength. The studies also show clearly that the size and shape of the target have a significant effect on the force exerted on it by the jet. A flat surface receives a greater force from the jet than a target with an inclined surface.

In late 2008, a survey on lime kiln operation and fuel usage was conducted through questionnaires distributed to pulp mills in nine countries. Responses were received from 59 pulp mills, totaling 67 lime kilns. This paper discusses the key findings from the survey with respect to kiln design parameters, operating data, control strategies, fuel types, energy consumption, and operating issues. Where possible, comparisons were made with findings obtained from a similar survey conducted in 1991. Over the past two decades many improvements have been made in kiln design and operation. Included are a larger number of kilns equipped with product coolers and lime mud dryers, along with changes designed to deliver higher mud dry solids content to the kiln. However, ring formation and dusting/high dust load remain the major operational issues among the respondents. Although few kilns have used alternative fuels to date, there is considerable industry interest in their use in the future. About two-thirds of the kilns have plans for implementing alternative fuels within the next five years.

The authors also wish to acknowledge Tero Karttunen and Markku Tanttu of YTI Research Centre for their assistance with probe design and construction of the probe, together with Lars-Gunnar Magnusson of Andritz, for their leading role in the collection of field data. Thanks also due to the SCA Ob - bola mill for providing the test site and help during the trials. This work was conducted as part of the research program on “Increasing Energy and Chemical Recovery Efficiency in the Kraft Process,” jointly supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) and a con - sortium of the following companies: AbitibiBowater, Alstom Power, Andritz, Babcock & Wilcox, Boise Paper Solutions, Carter Holt Harvey, Celulose Nipo-Brasileira, Clyde-Berge - mann, Diamond Power International, Domtar, DMI Peace River Pulp, Fabria, Georgia Pacific, International Paper, Irving Pulp & Paper, Metso Power, MeadWestvaco, StoraEnso Re - search, and Tembec.

The recovery boiler at the Irving Pulp and Paper Saint John, NB, mill has been through several major retrofits to increase its original firing capacity of 1100 metric tons/day of black liquor dry solids to the present level of 1680 metric tons/day. Many problems have been encountered over the years, including tube corrosion and cracking, as well as plugging of flue gas passages, but they all have been overcome through operational changes and process optimization. The latest challenge is to increase throughput without experiencing high total reduced sulfur (TRS) levels that would impact the environment and jeopardize compliance. An optimization program has been in place at the mill since December 2008 to further increase boiler production while maintaining environmental performance. The program, which automates liquor addition at full capacity according to targeted stack gas O2 and TRS levels, enables the boiler to operate at lower stack gas O2 targets and to achieve a 2%-3% increase in liquor throughput, while keeping TRS emissions under compliance.

Severe corrosion led to an in-service failure of one of the six spouts on a recovery boiler at Cenibra. The failure resulted in a large amount of cooling water entering the boiler and contacting smelt around the spout opening, but fortunately no smelt-water explosion occurred. An extensive investigation was conducted to determine the causes of the spout failure. The failure was caused by insufficient cooling water flow to the smelt spout, which, in turn, was caused by a sequence of events that occurred after a general blackout of an electrical energy system at the mill. The experience has led to the development and placement of operating and maintenance procedures that ensure adequate cooling water flows to smelt spouts to prevent future failures.

We examined the effectiveness of thermal, caustic, and sonication pretreatment methods in improving anaerobic conversion to biogas of secondary sludge samples obtained from a kraft mill and a sulfite mill. All three methods improved the anaerobic digestion rate and the biogas yield of the sludge samples. Thermal pretreatment was the most effective, followed closely by caustic pretreatment, and sonication the least. The total biogas productions per unit of chemical oxygen demand of sulfite sludge and kraft sludge samples were respectively 1.2 and 3 times higher with pretreatments than without. Also, the biogas production from the untreated sulfite mill sludge was 4 times higher than that from the untreated kraft mill sludge.

Failure of brittle solid bodies due to the impingement of a high velocity air jet on the body surface is studied, experimentally and theoretically. Using the linear elastic theory and stress distribution analysis, a general criterion for the failure of brittle materials impacted by a gas jet is derived. Several special cases of jet–solid body interaction including failure of thin and thick layers and cylindrical objects immersed in a crossflow gas stream are investigated and proper material failure criteria are developed. These criteria correlate the minimum jet peak impact pressure (PIP) required to break the material to the material's tensile strength and Poisson's ratio. A series of experiments were performed using a laboratory-scale apparatus. Gypsum cast on steel tubes forming cylindrical samples was used as the model brittle material. Experimental data and high-speed breakup movies are employed to understand the gas jet–solid body interaction and to validate the theoretical criteria developed for the material failure. It is deduced that the failure of cylindrical samples impacted by a gas jet is by the formation and propagation of cracks. However, when the impact jet diameter is small, the cracks cannot propagate, and the material is failed due to localized surface pitting. One of the practical applications of this research is in Kraft recovery boilers, where high velocity supersonic steam jets are employed to remove deposits accumulated on the outer surfaces of the steam tubes.

The solubility of recovery boiler precipitator ash in water is important in the design and operation of ash treatment systems used to control chloride (Cl) and potassium (K) accumulation in the kraft recovery cycle. We determined the solubility of mixtures of alkali salts that constitute precipitator ash and several precipitator ash sam-ples, through laboratory experiments and theoretical simulation using a commercially available thermodynamic pro- gram. The solubility of precipitator ash in water at 85°C varies widely, from 400–550 g/L H2O, depending on ash composition. The composition of the liquid portion of an ash-water slurry changes significantly with an increase in ash concentration: the Cl concentration increases, the sulfate (SO4) concentration decreases, and the K and carbon-ate (CO3) concentration remain the same. Between 30ºC–100ºC, temperature has little effect on the solubility of pre-cipitator ash and on the composition of the liquid portion of the slurry.

This work was part of the research program on Increasing Energy and Chemical Recovery Efficiency in the Kraft Pro - cess. The program is jointly supported by the Natural Scienc - es and Engineering Research Council of Canada (NSERC) and a consortium of the following companies: Abitibi-Bowater, Alstom Power, Andritz, Aracruz Celulose, Babcock & Wilcox, Boise, Celulose Nipo-Brasileira, Carter Holt Harvey, Clyde- Bergemann, Diamond Power International, Domtar, DMI Peace River Pulp, Georgia Pacific, International Paper, Irving Pulp & Paper, Metso Power, MeadWestvaco, Stora Enso Re - search, Tembec, and Votorantim Celulose e Papel.

The breakup by supersonic air jet of cylindrical brittle gypsum deposits formed on tube surfaces was studied using high-speed imaging. The breakup of symmetric deposits was visualized from the front and back simultaneously, and the effects of deposit asymmetry, jet/deposit attack angle, and jet duration and frequency, were investigated. Three jet types were considered: short and long duration single pulses, and a pulsating jet. For each experiment, the time to breakup, the breakup duration, and the breakup length were measured. Results indicate that the breakup of asymmetric deposits depends on orientation, and that attack angle affects the probability of breakup and breakup length. Results also indicate that soft deposits can be broken by any jet type, while harder deposits, especially those farther from the nozzle exit, require a longer exposure time to be broken. Pulsating jets do not improve breakup effectiveness. This work has relevance to cleaning of fireside deposits formed on heat exchanger tubes in boilers.

Many papers have reported that the standard k − e model fails to accurately predict the mean velocity profile of turbulent axisymmetric jets (Thies and Tam, Computation of turbulent axisymmetric and nonaxisymmetric jet flows using the K − e model, AIAA J., 1996, 34(2), 309–316; Pope, Turbulent Flows, 2002 (Cambridge University press: Cambridge). As the jet velocity increases, the deviation of the model with respect to the experimental measurements also increases. This work is aimed at the development of a modified k − e model that can be used to predict the mean properties of an axisymmetric jet as it (i) flows as a free jet, (ii) propagates between walls, and (iii) impinges on a solid object. Three additional terms are proposed to improve the standard k − e model predictions. They are Durbin realizable, Heinz turbulence production and Sarkar compressibility correction terms. The performance of the modified model in predicting the velocity and the impact pressure profiles of a free jet with an exit Mach n...

In this work, an advanced CFD-based ash deposition model is presented that allows for the quantitative prediction of the transport of fly ash particles near furnace walls and heat exchanger surfaces, combined with a detailed description of the thermochemical properties of the particles that consist of salt mixtures melting over a wide temperature range rather than at one specific melting point. The physical and chemical phenomena considered in this approach are: particle melting, particle impaction, particle sticking as a function of particle melt fraction, restriction of deposit layer growth by melt flow, and particle rebounding from the furnace wall or superheater surface. The model is validated with entrained flow reactor experiments at well-defined conditions. Three model substances consisting of mixtures of sulphates and chlorides are fired in the reactor with particle sizes typical for biomass combustion in fluidised bed combustors, ranging from smaller than 75 to 250 μm. The results indicate that while the proposed deposition model is capable of predicting deposit buildup on cooled surfaces, a clear distinction between physical and chemical effects contributing to deposit buildup for different particle size classes has to be made. For fly ash particles smaller 75 μm, all particles containing 15% molten phase or higher will stick to cooled surfaces as soon as they are transported to them and as long as the deposit does not grow above the steady state thickness. Larger particles may follow the same trend, but the high probability of them rebounding is a function of the impact angle and needs to be considered in the model.

The interaction between a mildly underexpanded supersonic jet and a single cylinder was studied experimentally at laboratory scale by using the schlieren technique coupled with high-speed photography and pitot pressure measurements. This study was motivated by the need to optimize sootblowing operation in kraft recovery boilers. The effects of the transverse distance between the jet and cylinder centerlines (eccentricity), nozzle–cylinder distance, and cylinder size on jet–cylinder interaction were determined. Results show that upon impingement on a cylinder, a supersonic jet deflects at an angle and creates a weaker supersonic jet that we refer to as a “secondary” jet. The angle and strength of the deflected or secondary jet depend on the eccentricity between the primary jet and cylinder centerlines. When a jet impinges on a cylinder of diameter comparable to that of the jet or smaller, secondary jets form not only when the cylinder is placed close to the nozzle (in the stronger portion of the jet) but also when the cylinder is placed far away (in the jet's weaker portion; up to 20–24 nozzle exit diameters in the present study). Changing the eccentricity slightly results in a significant change in the secondary jet characteristics. For a cylinder much larger than the jet, secondary jets do not form at zero eccentricity (head-on impingement); the eccentricity at which they begin to form increases with the cylinder size. A study of the secondary jets shows that they spread out much more than the primary jet and are sheet- or fan-like with an oblong, oval cross section. The centerline pitot pressure of the secondary jets remains as high as the primary jet for a considerable distance from the tube only during weak interaction between the primary jet and the cylinder (i.e., during strongly eccentric/off-centerd impingement). As the interaction between the primary jet and the cylinder intensifies at lower eccentricities, the maximum centerline pitot pressure of the secondary jet decreases, and the pitot pressure decreases more quickly with distance from the tube.

An explicit cell-centered finite volume solver coupled to a k−ε turbulence model corrected for structural compressibility fails to satisfactorily predict the behavior of under-expanded supersonic jets exhausting into still air, because the model does not properly account for the turbulence/shock wave interaction. Two approaches are examined: imposing a realizability constraint and taking into account shock unsteadiness effects. Although both corrections yield better agreement with experimental data of under-expanded jets, the realizability constraint yields better results than the shock unsteadiness correction.

An experimental investigation is conducted to determine the effect of liquid viscosity and density, nozzle diameter, and flow velocity on the break-up length and spreading angle of liquid sheets formed by splash plate nozzles. Various mixtures of corn syrup and water were used to obtain viscosities in the range of 1–170 mPa s. Four different splash plate nozzle diameters of 0.5 mm, 0.75 mm, 1 mm, and 2 mm, with a constant plate angle of 55 deg were tested. The liquid sheet angles and the break-up lengths were measured at various operating conditions. An empirical correlation for the sheet spreading angle and a semi-empirical correlation for the sheet break-up lengths are developed.

The failure of a cylindrical brittle material impacted by a supersonic air jet is investigated. Gypsum was cast around steel tubes to simulate the deposit formed on tube surfaces in industrial boilers. The breakup behavior of two deposit sizes, positioned at several distances from the nozzle exit, was visualized and documented using a high-speed video camera. Three deposit failure behaviors were observed: (i) crack formation and propagation along the longitudinal axis of the cylinder, (ii) surface pitting followed by axial crack formation, and (iii) surface pitting followed by spalling. These types of failure depend on the ratio of jet diameter to deposit diameter, which affects the magnitudes of compressive, tensile, and shear forces induced within the material. By analyzing the breakup movies, characteristics of the broken deposits, such as the breakup duration and the amount of deposit removed, were measured. Also, the effects of deposit thickness and distance from the nozzle exit on these characteristics were investigated.

The interaction of molten smelt and water inside a recovery boiler dissolving tank produces loud noise and can be violent even during normal boiler operation. Inadequate shattering of the smelt stream leads to even more violent interaction, as evidenced by an increased acoustic intensity of the dissolving tank. On rare occasions, a violent dissolving tank can explode, causing equipment damage and even injury to personnel. To warn operators of changes in dissolving tank conditions, an acoustics-based monitoring system could be developed. To assess the feasibility of such a system, acoustic observations were recorded at three pulp mills. Analysis of the recordings indicates that when a smelt stream is not being shattered, the intensity of the dissolving tank soundscape increases significantly and the frequency spectrum changes. We also observed a large variation between different mills both in average intensity and in signal variance. The results of this study suggest that the development of a monitoring system is feasible.

A scaled-down experimental apparatus was built to examine smelt shattering during typical recovery boiler operations. Water-glycerine solutions and air were used in place of smelt and steam. A high-speed camera and image processing software were used to record and quantify liquid shattering in terms of droplet number and size distributions, as a function of air velocity, air nozzle position, liquid flow rate, and liquid viscosity. The results showed that increasing shatter jet velocity reduced average droplet size, increasing the liquid flow rate increased droplet size, and placing the shatter jet nozzle closer to the liquid stream decreased droplet size. These results were all as expected. The effect of liquid viscosity (1-50 cP) depended on the shatter jet velocity. At high air velocities, even the viscous liquid was well shattered, but at lower velocities, the effect of viscosity on shattering was significant.

The authors express their appreciation to Metso Power Oy, the Finnish Foundation for Technology Promotion (TES), Kaartokulman erikoisrahasto, Walter Ahlstromin saatio, and the Finnish Graduate School in Computational Fluid Dynamics in the Academy of Finland for financing this project. They also gratefully acknowledge the contributions of J. Pyykonen, who proposed the principles on which the model is based, and of V.-P. Heikkila for developing the computational mesh and the combustion model.

A laboratory apparatus was constructed to simulate the operating conditions of recovery boiler smelt dissolving tanks and used to systematically study the interaction between molten smelt droplets and water. Experiments were performed on synthetic smelt made of 80 wt% Na2CO3 and 20 wt% NaCl at 800°C, 900°C, and 1000°C. The results show that upon contact with water, some smelt droplets explode immediately and break into small pieces, some require a delay time to explode, and others solidify without exploding. The probability of explosion strongly depends on water temperature and to some extent, smelt temperature. At a given smelt temperature, there exists a water temperature range below which explosion always occurs (the lower critical water temperature) and above which there is no explosion (the upper critical water temperature). The lower critical water temperature decreases with increasing smelt temperature, while the upper critical water temperature remains the same at 82°C in all cases. Up to this upper critical water temperature, both the explosion delay time and explosion intensity increase with increasing water temperature. The data was used to construct a Smelt-Water Interaction Temperature (SWIT) diagram that can predict if a molten synthetic smelt droplet will explode in water at different smelt and water temperatures.