"The current state of the global pulp and paper industry can be summed up in a few words: our profitability is an embarrassment to our management, a frustration to our employees, and a reason for investors not to buy our stock." Delegates at the 1998 International Chemical Recovery Conference, held in Tampa, FL, on June 1 to 4, 1998, received this message at the keynote address. Richard Phillips, International Paper, opened the conference with a reminder that the pulp and paper business has to make sense.
With a return on investment of 5% to 6% (for US companies, and about 3% to 4% for Canadian companies) the industry returns less to investors than do most risk-free government bonds. The Paper Industry Stock Index has grown at half the rate of all stocks over the past five years. Why? Well, Phillips proposed several reasons. First, the paper industry is the most capital intensive in the world, yet the rate of return is dismal. It is not a project management problem. "The industry is good at project management," he said. Engineering and construction is competent and, and there are no cost overruns. The problem, said Phillips, must be the projects that we choose.
Second, low cost producers in the Southern Hemisphere have eaten in to the markets of North American and Scandinavian producers.
Third, in the United States (and Canada) the forest products industry does not garner support from the public nor the government.
Fourth, the industry shares its technology and, therefore, no one has a competitive advantage, except for the suppliers, who will eventually market their innovations.
The first problem to tackle, and the most serious, is the poor choice of capital projects.
Choice of capital projects: What kinds of projects do make sense? "Paper consumption continues to grow. As developing nations become wealthier, they use more paper and packaging. Most of the paper producers can find clever ways to debottleneck their current operations by 1-2% per year, and thus, for the domestic markets for the developed nations, match the normalized growth curve." Phillips says that the most of the top five producers can also find ways to reduce their cost such that the decline in real prices can be matched with a decline in real costs.
Low-cost competition: In the Southern hemisphere, trees grow rapidly, and labor costs are low. Add to that the fact that most of the commodity paper products can be supplied by fibre that requires no recovery boiler or caustic plant at all. Phillips noted that the long-term erosion in the real price of products creates a very poor justification for recovery boiler replacement or modernization. He also pointed out that with no new recovery boiler market, there is little incentive for either the boiler manufacturers or the paper companies to invest in the R&D to make the processes more efficient in terms of either energy or in terms of capital dollars. His point: "No new technology, no need to have an International Chemical Recovery Conference. Hence, you become irrelevant."
Lack of support: The US does not add tariffs to paper imported from competitor nations. "Paper tariffs and other trade barriers are well down the scale on the US Trade Commission agenda for negotiation." This, according to Phillips, is due in part to the fact that "this industry has not effectively communicated its strategic importance to the general population."
Shared technology: Our industry is an exclusive club where the same small group of machinery and chemical suppliers provides the same technology. The problem, says Phillips, is that no supplier or mill gets competitive advantage, so if a shared technology lowers costs, for example, then all who use it will benefit from those same lower costs.
One answer to this global profitability crisis "is taking place in this room and outside this room," said Phillips. The recent trend toward mergers and acquisitions means fewer players with greater shares in the global market, and assets that don't leave the industry. These fewer players are better able to justify investments in new processes and new products.
Another solution would be the growth of paper demand. The paper consumption trend in Europe and North America has not been positive. A lot will depend on the consumption in Asia, China in particular.
The conference had 107 oral and 37 poster presentations, more than any other, according to program chair Mikko Hupa, bo Akademi University. Several sessions ran concurrently, and are not covered here.
A paper done by a group of researchers from Princeton University, Union Camp's R&D division, IPST and Weyerhaeuser looked at the economic value of a black liquor gas project. A Princeton University model of a power-house was combined with mass and energy balance process models of the pulping and chemical recovery areas. Four scenarios were considered.
Green liquor scrubbing: Increased caustic cycle costs;
Hydrogen sulphide gas recycle;
PS/AQ -- some benefit from increased yield, but less black liquor or less power also increased anthraquinone;
Air blown gasifier with hydrogen sulphide recycle saved $17 million more for gasifier. Substantially lower costs were achieved, but electrical benefit was not as great.
The development of a combined cycle is moving forward. This work shows that the increase in power production is a benefit of commercial application.
A spreadsheet is being developed to look at other integration scenarios.
Chris Gore, Willamette Industries, decribed how his company increased black liquour processing capacity at its Marlboro, SC, mill, as well as maximized capacity of RDH bleached pulp mill. The evaporators were modified to increase solids concentration as well as increase capacity. The evaporator was upgraded with six falling film units with intermediate soap skimming. The new flow sheme downgraded evaporator 1 plus concentrator 1 to 50% final product solids and its product rerouted to a coned bottom 50% tank. All of the combined liquor to the two evaporator trains were pumped to concentrator 2 and evaporated to 80%. This 80% product liquor were stored in pressurized tank with coned bottom and recirculation. Boiler ash and purchased saltcake were picked up with a side stream to and from the 50% tank and blended into the feed liquor to concentrator 1.
Christopher Beveridge, ABB Power Generation, discussed some R&D developments in non-cooled smelt spouts, particularly a spout suitable for use on a kraft or soda recovery boiler. He discussed six different mill installations across North America and thechallenges that they have had with corrosion and cracking. According to Beveridge, the new cast alloy used exclusively for non-cooled spouts fabricated in the past two years has exceeded expectations with respect to its resistance to wear.
Hirotoshi Matsumoto, Mitsubishi Heavy Industries, stuck to this corrosion-combatting topic, describing advanced technology for corrosion resistant materials for recovery boilers. This paper supported the points that Doug Singbeil brought up in his round-up presentation of the corrosion conference. (Refer to PULP & PAPER CANADA 99:12.) Corrosion tests at 400C showed that specially developed 25% Cr overlaid material had good corrosion resistance in both reducing and oxidizing atmospheres. An 18% Cr overlaid material was found to have good corrosion resistance in a reducing atmosphere at 400C. Cr was effective in preventing corrosion in the environment of recovery boilers. Materials including 18% Cr and over have excellent corrosion resistance in both reducing (1%H2S) and oxidizing (1% SO2) atmospheres. Corrosion tests simulating the environment of the superheater showed Cl induced intergranular corrosion is seen to be caused by Cl2 gas formed by the reaction between ash and atmosphere when the melting of ash starts.
Extensive mill sampling was used to develop mass balance process models, in a study done by Murray Ellis, PAPRO, New Zealand. Approximately 100 samples were done. Three of four samples were taken from every point throughout out the day. A similar study was done here in Canada by Paprican. The mass balances around the kraft pulping showed that sodium and sulphur behaved similarly during pulping, leaving the fibreline almost entirely with the black liquor.
Pulp leaving the unbleached fibreline acted as a silicone purge for the liquor circuit while the bleach plant produced pulp containing more silicone than before bleaching.
In the question period Ellis said that silicone and iron were the two most difficult elements to evaluate.
James Frederick, IPST, presented a paper entitled The solubility of aluminosilicates in kraft green and white liquors. The study sought to determine the solubility of aluminosilicates and develop strategies for their effective removal.
Aluminum and silicone can be removed effectively from green liquor by precipitation with magnesium salt, and efficient dregs removal. The research found that the concentration of silicone in the clarified liquor can be reduced to less than 100 ppm. Aluninum can be reduced to at least 100 ppm and perhaps below 10 ppm depending on the ratio of Al/Si in the smelt entering the tank. Effective dregs removal and good lime quality are essential to prevent Al and Si in white liquor.
The strategy adopted: Acceptable Al and Si levels in clear green liquor depend on input between green liquor clarifier and black liquor evaporator. Potential sources include lime, wood chips defoamers, and white water. Mass balances will determine what is the acceptable level in green liquor.
Michael Paleologou, Paprican, spoke about the precipitator dust purification (PDP) system that was developed by Paprican and Prosep researchers. The PDP addresses high levels of chloride that can lower the range of temperature at which sticky deposits are produced on the superheater and the generating tank tube of the recovery boiler. This leads to plugging of the flue gas passages. The PDP system removes sodium chloride from an aqueous solution of the electrostatic precipitator dust of the recovery boiler.
Benefits of the PDP include:
High chloride removal efficiency;
Low capital and operating cost;
Low energy requirements; and
Easy integration into the recovery cycle.
The burning bed
The recovery boiler modelling session started out with a discussion on high solids firing in an operating recovery boiler. Esa Vakkilainen, Ahlstrom Machinery, described how a three-dimensional flow simulation was applied to analyze the high solids firing process in a 3150 tds/d recovery boiler. The flow field, combustion, heat transfer, and the concentration of nitric oxide (NO) were modelled with the commercial Fluent/UNS, notes Vakkilainen.
A strong local refinement, he said, of the computational grid was used in the parts of the furnace with fuel and air injections and close to the char bed. The trajectories of the black liquor droplets were computed with a stochastic lagrangian method.
Results from the calculations were compared to measurements. Conclusions showed:
N release agrees with measured values;
Predicted final level lower than measured level;
Odorous gas compounds can affect the final NO level.
Tom Grace, IPST, presented two papers consecutively. The first discussed a new black liquor drop burning model which: can permit rigorous mass and energy balances; develop a model for burning on a char bed without making a lot of changes; has the capability to predict air emissions; and the capability to predict fume formation. "The model," the authors note, "allows computation of parameters that relate directly to boiler fouling and plugging. It also allows the prediction of the concentrations of air emissions including TRS, CO, VOC, SO2, and NOx.
"The second paper also deals with the DOE model," said Grace. "It is a validation of the CFD-based recovery furnace models." The model validation is described briefly in the preprint and in detail in the DOE final report. It consists of validations of: flow predictions; black liquor combustion models; complete recovery boiler model; and code benchmarking.
IPST and Radian International worked on a project dealing with artificial neural network modelling of a kraft recovery boiler. The study was done to discover whether neural network models could be used to describe the recovery boiler operation. They developed control models. The neural network model, said Mark Denlinger, Radian International, can produce accurate predictive models to predict O2, CO2 and NOX emissions. It is, he said, a good tool for large data sets with a lot of noise present.
Another three-dimensional computer model was developed, this time to simulate the flow, heat transfer, combustion and NOX emission in a kraft recovery boiler. The model was used in a study done by the Royal Institute of Technology in Stockholm, Sweden, to evaluate the combustion and NOX control alternatives in a utility kraft recovery boiler in Sweden. The boiler used had an air system featuring a rotation firing configuration in the lower part of the furnace and double overfire air in the upper part. The recovery boiler model was developed using a general purpose Computational Fluid Dynamics (CFD) package. To overcome the numerical difficulties experienced using a structured computational grid, due to the complexity of the flow and geometry, an unstructured grid and with local web refinement was used. This modification made it possible to simulate the complex flow and geometry of a kraft recovery boiler with realistic computing power.
Operation and design
Andrew Jones, ABB Power Generation, presented a paper that discussed the claims made for high solids firing. A performance test was conducted on the SAPPI-S.D. Warren kraft recovery boiler in Westbrook, ME. The solids levels tested were 71%, 76%, and 81%. Jones discussed the likely and possible effects of firing percent dry solids in 5% increments. Some of these are: increased NOX emissions; reduced carryover; lower sulphur; small increase in dust emissions from boiler; did not see a difference in pumping high solids liquor; no char bed control problems.
Kaj Backman, Stora Cell, described experiences with new technology at the Stora Cell Skutskr, Sweden, mill. This presentation summarized the major modifications done in the recovery system of the Skutskr mill. A vertical air system in recovery boiler 7, burning NCGs in recovery boiler 6 and a new evaporator line with an integrated stripper are some of the modifications made. Extended condensate segregation together with extensive condensate stripping reduces the COD load in the mill effluents by more than 10 kg/adt. The vertical air system implemented on RB7 has allowed an increase in burning capacity by 25%, to above 2400 tds/d. The production targets: from 430 000 to 550 000 t/y by the year 2000. To meet these targets for production dry solids need to increase from 2400 to 3100 tds/d; white liquor increase from 5500 to 7000 m3/d; and, evaporation from 500 to 670 t/h. A training program has also been put in place to support the continuous improvements. At the time this paper was presented, the mill reported two months of undisturbed operation. "Model prediction control has given excellent results in stabilizing the front and back end temperature of the kiln.
Another paper looked at the practical experience of modernizing a two drum recovery boiler to a single drum type. "A single drum conversion gives an old recovery boiler a second life," said Kjell Ljungvist, Kvaerner Pulping. "Why convert a two-drum boiler to a single drum?" Ljungvist answered his own question by listing the benefits that were seen in four cases (two in Sweden, one in both Germany and Norway). The single drum designs combined with cross-flow panel-type generating banks showed:
Improved safety "because the system works according to the laws of nature". The drum design increases the overall circulation in the boiler, which means better cooling of the furnace water tubes;
Less maintenance and therefore reduced costs;
Jan-Erik Gustafsson, STFI, presented a study that measured residence time distribution in a recovery boiler "with high accuracy". The method developed uses helium injection in the combustion air. Measuring the resulting helium concentration, the residence time distribution of the recovery boiler can be determined. Gustafsson said that the work determined the residence time of the complete system. The work was compared with CFD-model, but did not correlate. The authors believe that this is due to the fact that the mixing is not described correctly in the CFD-model.
Tom Grace was up again and this time talked about avoiding an explosion in the thermal bed. A recovery boiler emergency shutdown procedure (ESP) has been developed. He said that this topic is also being study by AF&PA and others.
All actions taken after an ESP should not increase the extent of combustion on the bed. This paper examines the procedure for measuring temperatures in char beds and for using such temperature data in assuring the thermal state of the cooling bed. Three methods were used to monitor the beds: thermocouple probes; floor tube thermocouples and IR scanning of the bed surface. The conclusions showed that continued combustion is a critical factor, as it increases the cooling load. Enhancement of thermal conduction within the bed with interpore radiation is also an important factor. Measurements of bed surface temperatures are not a good indication of the thermal state of the bed underneath. Thermocouple probes need to be two to three feet into the bed to provide a reasonable measure of the internal bed temperature. "Thermocouples have to be below the bed's surface," said Grace. "Get the probes in early, before the bed freezes up. Drive them in if you have to."
In session 10, Causticizing, J. Kivivasara, UPM-Kymmene, performed an evaluation of various alternatives to reduce TRS emissions of the lime kiln. In this field survey, lime mud filtering and lime kiln operation were studied to discover the influence of operational parameters on the TRS emission. A neural network was developed to predict TRS emission.
"Using a neural network, and on-line data, predictions can be reasonably good." The results showed that the lime kiln TRS emissions are directly related to the function of the lime mud washing and filtration operation, and lime kiln burning conditions.
A high level control system, the FuzzyExpert, was installed at the Stora Skoghall, Sweden, mill to control the causticizing plant. Relevant measurements are input into the expert system, which evaluates the information and calculates the ideal set-points for the control of the three sub-processes; lime reburning kiln, slaker, and causticizer. The expert system keeps the process at its optimum and attains a stable lime quality, optimizes production span, and reduces fuel consumption.
The results obtained were beyond expectations, said Jeffrey Kemmerer, with the most dramatic improvements around the kiln. He saw a reduction in fuel consumption, expanded production range, more stable and improved lime quality, more reliable kiln operation, and more stable white liquor quality as measured by effective alkali.
He cautioned, however, that the system is user sensitive, and great care should be taken in training operators. "If the operators don't like it, if it can't be easily maintained, if no one understands it, (or) if those responsible for it don't trust it, it will never work." He also suggested that mills dedicate one process engineer to maintain and update the systems when reconstruction takes place or when conditions change.
In order to prepare Aracruz Cellulosa, Brazil for water reduction measures and future effluent closure the mill underwent a modernization program. Pulp production capacity was increased to 1.24 million tons of bleached eucalyptus pulp per year. A green liquor filtration plant was installed. Six months after the start-up significant reductions of dregs and non-process elements in the liquor cycle were shown. Metals removal efficiencies decreased. Both the green and white liquor improved. From the environmental point of view, the authors note, the significant reduction of metals in the liquor is an important step to reduce non process elements in the pulp production process and contributes with the closed loop mill operation.
The next paper presented in theis session on causticizing measured the strength of limes from fluidized beds and rotary kilns. The study aimed to discover why limes from fluidized beds give increased strength and specific surface area than limes from rotary kilns. Josefina Lindblom, Chalmers University of Technology, notes that when studying formed lime pellets reburned in a laboratory furnace, it was shown that all the investigated parameters, such as compact pressure, temperature, and sintering time, influenced the sintering process. Increasing temperature and sintering time results in a smaller specific surface area, whereas a higher compact pressure gives a larger specific surface area. The author also notes that with increasing compact pressure, the strength of the sintered agglomerates increased. When the lime was calcined before being pressed into pellets instead of afterwards, the strength became much higher. The results may explain the harder agglomerates from fluidized beds, as well as the hard shell often found on agglomerates from the rotary kiln.
Barry Hirtz, Canadian Forest Products, presented an advanced control strategy, which reduced recovery boiler fouling while liquor throughput increased. Hirtz reported that liquor solids throughput on the Intercontinental Pulp mill boiler, in Prince George, BC, increased 8%, on average, from 1460 t/d to 1580 t/d of dry solids. Recovery boiler operation on advanced control has resulting in stabilizing furnace conditions. There has been a significant decrease in the rate at which ash builds up on boiler convection surfaces. Boiler thermal efficiency has improved through a reduction in the excess air for combustion, with the oxygen in the flue gases reduced from an average of 2.8% to 1.5%.
Adhesive property of recovery boiler ash is influenced by potassium and chlorine concentrations. By decreasing K and Cl adhesive property is improved. A study done in Japan developed potassium removal equipment that removes K and Cl ash collected. "We made long term and continuous generation of the recovery boiler possible by adopting the potassium removal equipment, in order to improve adhesive property of ash," reported Masakazu Tateishi, Mitsubishi Heavy Industries. Ash is collected with an electrostatic precipitator. Water is mixed at a temperature no more than 30C. K2SO4 and NaCl are dissolved in the water when cooled: Na2SO4 precipitates as Na2SO410 H2O crystal. By separating this crystal and solution, K and Cl are removed simultaneously and Na2SO4 can be collected (for cooking).
In another study, a laboratory scale char bed reactor at IPST was reconstructed to study the formation of particulate during the char burning stage of black liquor combustion. The reactor was connected to a particulate collection system. Three char samples were used with one black liquor source. Two types of particles were observed after char burning: submicron fume and 1 to 100 m. The larger particles are probably formed by mechanical ejection of char bed materials, according to Saied Kochesfahani, University of Toronto. These particles are called ejecta. The study characterized and quantified particulates formed in char burning, determined the quantity and composition of these particles under controlled conditions and , determined the effects of operating variables on the quantity and chemistry for the particulates.
Submicron fume particles and 1-100 m ejecta particles form in large quantities during char bed bu rning experiments. More than 90% of the ejecta have a size in the range of 5 to 40m, with an average of about 20. "About 40 to 85% of the total particulate formed is ejecta," said Kochesfahani. "The balance is fume." Kochesfahani gave the following summary:
The quantity of ejecta increases markedly with temperatures up to 1000(C, but reaches an almost constant value at higher temperatures.
The carbonate content of fume increases with the bed temperature, while the sulphate, chloride and potassium contents decrease.
The composition of ejecta was found to be similar to oxidized smelt and is consistent with mechanical mechanism of formation.
To gain better prediction of the liquor behavior in combustion or gasification processes, work at bo Akademi has tested different mill liquors. A method for predicting the composition of melting properties of the dust, which would be produced from firing liquor under typical furnace conditions, was developed. "The prediction method is based on liquor elemental analysis, single droplet pyrolysis tests, and advanced chemical equilibrium calculations," the authors note. The predictions are made for both "hot" and "cool" furnace conditions. The behavior of the sulphur and chlorine are the main difference between these two.
The results obtained were compared with dust data from the boilers where these liquors were burned. The predictions and comparisons with the dust sample data gave very encouraging results.
Another study, presented by Kauko Janka, Kvaerner Pulping, looked at fouling and plugging of recovery boilers. The effects of black liquor composition and boiler combustion were studied, and the use of the enrichment factor concept as a tool for addressing fouling problems was further developed. "Enrichment factors are defined for samples collected from different sampling locations in the boilers and the role of furnace heat load, temperature and dust load on the melting temperatures T0 (first melting temperature), T15 (sticky temperature, and T10 (flow temperature) of the deposits, and on the superheater deposit thickness.... For the measured boilers, the effects of variations in furnace conditions on the first melting T0 are in the range of 10 to 20C. This is the variation of maximum safe superheating temperature, if the deposit smelt corrosion is the limiting factor. The total variation of maximum safe superheating temperature, if the deposit smelt corrosion is the limiting factor. The total variation (including the influence of different liquor) is about 40C. The sticky temperature T15 variation due to the furnace conditions is in the range of 100C, which is the same order of magnitude as the variation of the measured boilers is more than 200C." The authors also concluded that the furnace conditions do not affect significantly the flow temperature T70 or the deposit equilibrium thickness of the superheater deposit.
The next paper presented in this plugging prevention session explained how deposits are removed from kraft recovery boilers. The mechanics of deposit removal in kraft recovery boilers, presented by Don Cormack, University of Toronto, described sootblower hydrodynamics, the impact of jets on deposit and deposit strength.
"The role of full expansion nozzles are worth discussing," Cormack said. Full expansion nozzles reduce steam requirement for cleaning. They are designed so that the jet achieves full expansion before it exits the nozzle. This increases the jet energy available for deposit removal. At a given steam flow rate, full-expansion nozzles deliver high peak impact pressure (PIP), the dynamic pressure of the jet as measured by a pitot tube, than under expansion nozzles. One problem with the full expansion nozzles is that, because of their extremely high peak impact pressure (PiP) they can cause tube erosion.
The paper's conclusions show that:
calculations suggest that full-expansion nozzles, operated at 80% steam consumption rate, can achieve the same cleaning radius as an under-expansion nozzle under a range of boiler conditions;
deposit removal by debonding may be dominant in many locations in a recovery boiler, because for a sintered deposit, the adhesion strength is typically several times smaller than the deposit strength.
Cormack said that they have gained an "understanding of steam savings that arise from full-expansion nozzles", and have made significant progress toward the prediction of deposit removal by debonding.