SciELO - Scientific Electronic Library Online

 
vol.6 número2The release of Hydrocarbons from softwood drying: Measurement and modelingIDENTIFICATION OF WOOD DESTRUCTION DURING DRYING índice de autoresíndice de assuntospesquisa de artigos
Home Pagelista alfabética de periódicos  

Serviços Personalizados

Journal

Artigo

Indicadores

Links relacionados

Compartilhar


Maderas. Ciencia y tecnología

versão On-line ISSN 0718-221X

Maderas, Cienc. tecnol. v.6 n.2 Concepción  2004

http://dx.doi.org/10.4067/S0718-221X2004000200003 

Final Batch Average Moisture Content (%)

-

17.2

20.6

20.7

Average COP* (-)

HP-1

4.23

4.6

3.46

HP-2

3.70

4.07

3.00

Average SMER** (kgwater/kWh)

HP-1

2.38

2.13

1.46

HP-2

2.52

2.36

1.54

Average SEC*** (kWh/kgwater )

HP-1

0.42

0.47

0.68

HP-2

0.40

0.42

0.64

HP – Heat Pump; COP* – Coefficient of Performance; SMER** – Specific Moisture Extraction Rate (based on compressor and blower energy consumption); SEC*** – Specific Energy Consumption (based on compressor and blower energy consumption).

Operating Lessons Learned

Spread over several months, the first development step of high-temperature heat pump prototypes was aimed at ensuring a maximum operational stability of the thermodynamic parameters, establishing the reliability of the most sensible components (compressors, blowers, safety valves), checking the refrigerant/oil blend’s behavior and optimizing the critical control sequences of the system. The experimental dehumidifier dryer operated in “extreme” temperature, humidity and corrosion conditions, and demonstrated specific feedback phenomena that do not occur in traditional air-forced/heated lumber dryers. Because some of the by-products that could be produced by chemical interactions between the lubricant and working fluid may be acidic and lead to accelerating the corrosion of the system components, periodical controls aimed to determine the chemical behavior of the mixture. After about 3,250 hours of operation, the refrigerant proved to be thermally stable and chemically inert at the highest temperatures occurring in the system and a first oil chemical analysis proved that there were no problems with the oil breaking down or failing. The oil still showed adequate viscosity and chemical stability as well as a good miscibility with the refrigerant. The initial designed capacity of the heat pumps proved to be too high and consequently, both compressors were slowed down by about 25% which finally resulted in a more adequate capacity, reduced head pressures and improved efficiency. Because the original employed expansion valves poorly controlled the refrigerant flowing, did not open fast enough and manufacturer leakage faults were detected, they were later replaced by new generation devices. A crack in the casting of an original pressure relief valve also indicated a manufacturer’s defect, and finally both these components have been replaced to 20% higher pressure-limit valves. Other issues included the fact that the kiln was initially poorly insulated and leaky, and some system’s components were prematurely corroding.

CONCLUSIONS

As a clean energy technology compared with traditional heat-and-vent dryers, high-temperature heat pump dehumidifiers offer interesting benefits for drying resinous timber. This paper presents the preliminary results of the development and field testing of two prototypes accentuating their thermodynamic parameters, preliminary energy performances and first operating lessons learned. The average measured specific moisture extraction rate of the heat pumps was 2.35 kgwater/kWh (white spruce) and 1.5 kgwater/kWh (balsam fir), while the average coefficients of performance generally varied from 3.0 to maximum of 4.6. The cycle’s duration ranged from 2.5 days (white spruce) to 6.3 days (balsam fir) including the initial preheating steps. The refrigerant/oil mixture behaved well during more than 3250 hours of preliminary tests, proving good compatibility and chemical stability at condensing temperatures below 110°C (230°F). Better insulated and well maintained dryers are necessary to obtain drying temperatures higher than 100°C (212°F) as well as reducing the drying duration of resinous species by up to 25% and the total energy consumption by up to 50%. The current goals of the study include using more corrosive resistant components, variable speed central fan, further optimizing the drying schedules and general dryer operation and maintenance. Finally, it is expected to help local Canadian equipment suppliers to promote research and development of the technology and develop an appropriate market strategy. Specifications of high-temperature heat pump dehumidifier kiln energy use and a best-practice guideline must also be produced.

ACKNOWLEDGEMENTS

The author gratefully acknowledges “Hydro-Quebec’s Energy End-Use Service” for their indefectible support in this R&D project and our technician, Mr. Marcel Déry, who installed the monitoring system, helped in the long-term survey and contributed to the enhancement of the system’s electronic controls. Finally, the availability of the heat pumps, their field development and comprehensive testing were made possible thanks to a Canadian sawmill and to a North American heat pump manufacturer.

NOTE

♣ This paper was first presented at the IDS-2004, Sao Paulo and pre-selected for MADERAS:Ciencia y Tecnología journal.

LITERATURE

CANADA STATISTICS AND INDUSTRY MINISTER. 2002. Public Documents        [ Links ]

CECH, M.J; PFAFF, F. 2000. Operator Wood Drier Handbook for East of Canada. Edited by Forintek Corp., Canada’s Eastern Forester Products Laboratory        [ Links ]

KASACHKI, G.S.; GAGE, C.L; HENDRIKS, R.V. 1994. Investigation of HFC-236ea and HFC-236fa as CFC-114 Replacements in High-Temperature Heat Pumps. CFC’s: The Day After, Padua, Italy[         [ Links ]STANDARDIZEDENDPARAG]

^rND^1A01^nSTEFAN J^sKOWALSKI^rND^1A01^nANNA^sSMOCZKIEWICZ^rND^1A01^nSTEFAN J^sKOWALSKI^rND^1A01^nANNA^sSMOCZKIEWICZ^rND^1A01^nStefan J^sKowalski^rND^1A01^nAnna^sSmoczkiewicz

Maderas. Ciencia y tecnología, 6(2): 133-143, 2004

ARTICULO

IDENTIFICATION OF WOOD DESTRUCTION DURING DRYING♣

STEFAN J. KOWALSKI1, ANNA SMOCZKIEWICZ1
1Poznan University of Technology, Institute of Technology and Chemical Engineering, Pl. Marii Sklodowskiej Curie 2, 60-965 Poznan, Poland.

Corresponding author:
Stefan.J.Kowalski@put.poznan.pl, msmoczkiewicz@optimedia.pl

Received: 13.09.2004. Accepted: 24.11.2004


ABSTRACT

The subject of this paper concerns destruction of timber during drying. The main goal is to propose a method of avoiding destruction through a suitable programming of drying processes, controlled with the help of the acoustic emission (AE) method. Three different programs of convective drying of pinewood (Pinus sp.) samples are presented. The high and slow rate drying programs were applied to show an evident dependence between the intensity of AE signals (their number and energy) and the degree of destruction of pinewood during drying. The third drying program was controlled, i.e. the drying was accelerated, when the acoustic emission was low, or slowed down, when the acoustic emission started to grow rapidly. In this way , the drying process was optimized for the purpose of shortening of the drying time and avoiding a destruction of the material.

Keywords: destruction, pinewood, controlled process, acoustic emission.

 

INTRODUCTION

Drying of timber is the one of fundamental technological processes in wood industry. A proper realization of this process is necessary to obtain a good quality product. The high costs of drying involves to look for optimized processes with respect to drying time and energy consumption and, in particular, for processes not causing destruction of timber during drying. A very helpful in this afford seems to be the acoustic emission (AE) method that enable monitoring on line the material destruction, and thus, give a basis to drying control.

The studies on AE in drying of wood were already carried by some authors, e.g. Cunderlik et al. (1996), Kagawa et al. (1980), Kitayama et al. (1985), Noguchi et al. (1980), Skaar et al. (1980), Kowalski et al. (2004). These authors noticed that an increase of drying rate involves an instantaneous increase of AE signals originated from created defects in wood and the drying induced stresses. Therefore, one came to the idea to use the AE for monitoring the process and control followed by change in drying parameters. This is just the main goal of the present studies.

Monitoring a drying process by making use of the AE method one can observe the mechanical behavior of the material (in our case pinewood samples) during drying and thus to draw some conclusion from these observations for the purpose of construction of optimized processes. To this aim a number of tests in different drying conditions were carried out, starting from the high drying rates, at which an intensive cracking of wood structure was observed, up to the very slow drying rates, at which no any crack occurred. Based on the conclusions drawn from those studies, an optimized process was programmed with controlled temperature of the drying medium. The rate of drying in that process was changed periodically, i.e. it was increased at that moment, when the intensity of AE was low, and decreased, when the number of AE signals and, in particular, their energy started to increase rapidly. In such a way the time of drying was shortened by no cracks on the dried samples occured.

EXPERIMENTAL PROGRAM

The experimental equipment applied in our studies enables direct measurement of drying parameters (temperature and humidity of the drying medium), loss of sample weight, and in particular the AE signals, among others such descriptors as: the number of AE signals, the energy of AE signals, the total number of AE signals, the total amount of AE energy emitted. Figure 1 presents the drying set-up, with the measuring set for AE.

Figure 1: Measuring set for AE: : 1 - chamber drier; 2-balance;3-temperature and humidity sensors; 4-piezoelectric transducer;5-wood sample; 6-detector AE; 7-amplifier; 8-conversion of AE signals; 9-GPID card.


The acoustic emission (AE) method enables registration of sonic signals generated by fracture and microcracking of the structure in stressed materials. It provides unique advantages of early detection of crack growth and recognizing when the crack occurred. The piezoelectric transducer 4 can detect and a very sensitive receiver 6 can register, the acoustic signals transported by elastic waves released due to destruction of the material. The pinewood sample 5, placed in the laboratory dryer 1 constituted the generator of acoustic signals during drying. The transducer transforms the acoustic signals into electric impulses. These impulses arrive to the preamplifier 7 that has the following function: pre-amplifying the impulses and filter them with respect to frequency ranges (linear band, lower and upper band), and finally convert the voltage impulses into the current ones. The next element is the logarithmic converter 8 integrated with the pre-amplifier, which converts the current impulses of the transmission line into the variable voltage proportionally to the magnitude of a impulse, and next detects and enhances unipolar output signals. It is the peak detection of the envelope type from the absolute value of both the signal amplified logarithmically and that of the impulse type from maximal amplitude and the length corresponding to the time of signal duration. Both signals, the envelope (logarithmic) type and the impulse type, can be observed during the process on the oscilloscope HP 54603B.

All incoming data were possible to analyze on-line and store in the computer memory. The computer by way of a software compatible with the canvassing card GPIB 9 registered the acoustic signals and the remaining data, as the laboratory balance 2, temperature and humidity sensors 3 were also connected to the computer via GPIB card.

The pinewood sample 5 and the piezoelectric transducer 4 were suspended to the balance and weighted in prescribed time periods with accuracy up to 0.01 g. The loss of sample weight registered in time allowed to draw the drying curve. This curve enabled identification of AE intensity in time of drying and point out those moments, at which the AE start to increase rapidly.

The tested material was pine timber (Pinus sp.). The samples were cut out from a cylindrical timber having the wood core in the middle, and prepared in the form of cylindrical samples of 5 to 6 cm in diameter and of 1.5 to 2.5 cm height. The samples contained 10 to 12 annual rings. Their initial moisture content was c.a. 60%. Before drying they were wrapped up in foil to protect against the lost of humidity and stored in the laboratory cooler in the temperature 5 oC.

Drying of pinewood samples was carried out in the laboratory dryer in different temperatures from 80 to 100 oC. The metal plates of the sample grip covered the end faces of the cylindrical samples, so that the moisture removal proceeded through the lateral surface of the cylinder only. Choose of cylindrically shaped samples followed from the fact that wood suffers the most shrinkage in the direction tangential to the annual rings, what by covered end faces ensured enhanced AE. In order to ensure a good contact of the sample with the piezoelectric transducer, the bottom end face of the cylinder was greased with silica gel, and the sample was pressed to the transducer with an elastic element.

 

EXPERIMENTAL STUDIES AND RESULTS

High drying rate tests

The first series of drying tests were carried out in convective laboratory chamber at temperature of 100 oC and 4% humidity of the drying medium. Figure 2 presents the typical history of the drying medium temperature during these processes.

Figure 2: The temperature of the drying medium in the dryer by high drying rate.

The small drops of the temperature visible in the c.a. 230 and 270 minutes drying process are caused by open dryer door for short time. Figure 3 presents the history of AE events intensity, i.e. the number of acoustic signals per 30 s time intervals. Figure 4 presents the total energy emitted by the pinewood sample dried at 100 oC.

Figure 3: History of AE events intensity in pinewood samples dried at 100 oC.


Figure 4: Total energy emitted by the pinewood sample during drying at 100 oC .

 

Analyzing the obtained results of AE on tested samples one can differentiate three stages, during which the picture of recorded acoustic emission changed in time, and thus also the mechanical state of the samples. At the initial stage of drying, i.e. when the free water evaporates from the wet surface, the number of AE events is small, even less visible in the figure 3 up to 50 min drying time. Next, as the drying proceeds further, the acoustic activity starts grow intensively, and after 2.5 hour of drying time the number of AE events reach maximal value. The reason for the constantly increasing number of AE events in this stage is the shrinkage of external surfaces of the samples. When wood dries, the drier surface attempts to shrink but is restrained by the wet core. The surface is stressed in tension and the core in compression. The non-uniform shrinkage of wood tissue generates the stresses that cause microcracks at the surface, which number grow more and more if drying advances, (see Kowalski, 2003). The energy of the first AE signals is insignificant up to 100 min drying time. Since that time the energy start to grow rapidly.

The tensional stresses in the external layer of the cylindrical samples, when they reach critical values (yield stress), create macrocracks that are visible even with a naked eye. The creation of macrocracks is accompanied by the release of big energy. The portions of this energy are visible on the curve of total energy (Figure 4) as the vertical lines, the length of which depends just on the energy released. Figure 5 illustrate the macrocracks arose in the pinewood samples during high drying rate.

Figure 5: Appearance of macrocracks in pinewood sample arose during high drying rate.


In the series of 5 pinewood samples dried in high drying rate conditions one could observe from one to several macrocracks on the sample cross-section. Note that after two strong macrocracks the sample became unstressed, at least for some time, and the total energy grew no more since 290 min drying time.

Slow drying rate tests

The next series of drying tests was carried out at lower temperature of the drying medium, that is, at the temperature by which no any crack occurred in the dried sample. In our case it was c.a. 80 oC, (Figure 6).

Figure 6: The temperature of the drying medium in the dryer by slow drying rate.

 

Lowering of the drying medium temperature to 80 oC resulted in different picture of the AE. The shape of the curve presenting the intensity of AE events in this case (Figure 7) differentiate not so much from that by high drying rate at the first stage of drying, but the number of AE events decreased significantly.

Figure 7: Total number of AE events by slow drying rate.

In the case of high drying rate the maximal number of events reached 4000, while by the slow drying rate only 1000. Besides, in the first case the number of AE events dropped almost to zero after 200 min drying time, while in the second case not. This can be explained by the fact that the macrocracks in the first case caused the sample unstressed due to release of the cumulated energy. In the second case, on the other hand, no cracks occurred so that the sample was constantly slightly stressed. Therefore, the acoustics signals and the acoustic energy were emitted further in this case, however, with decreasing tendency as it concerns the AE events, and without rapid growth in energy (no vertical lines on the curve of total energy in figure 8).

Cracking in material structure, when they occur, manifest themselves on the total energy curves through the rapid increase of this energy.

Figure 8: Total energy of AE impulses by slow drying rate.

Controlled drying process

The characteristic feature of the acoustic emission method is that it provides a possibility of early detection of subcritical crack growth. The conclusions followed from the tests described above supply arguments to draw a controlled drying process, by which no cracks in the dried material occur and the drying time is relatively short. The experiment with controlled process was carried out in such a way that watching constantly the changes in AE events and the emitted energy one reacts on their changes by slowing down the drying rate when the AE increases or accelerate the drying rate when the AE is small. In our experiment the control was executed through alteration of temperature, as it shows figure 9.

Figure 9: The history of temperature of the drying medium in the controlled drying process.

 

Recalling the former experiments carried out in high drying rates, i.e. in the temperatures of 100 oC, one can state that the risk of cracking is not great at the initial period of drying. Based on this statement, the controlled process was begun at temperature of 100 oC. The development of AE was observed constantly and at the moment when the energy of the acoustic signals began to increase, c.a. at 150 min drying time, the temperature of the drying medium was decreased. It is worth to noise that when the energy started to increase the number of AE events slowed down. It means that the descriptor of average and total energy is the most relevant for drying process control.

Figure 10: The history of the AE events in the controlled drying process

The results of our studies point out that the fractures in pinewood structure may arise when a rapid increase of emitted energy take place. Only the AE signals originated from crack growth are recognized as being high energetic. The current control of the emitted AE signals, and in particular of the average and total energy emitted during drying, suggested the suitable changes in drying medium temperature in order to assure a reasonable dynamics of drying in order to avoid fractures of the material. The gentle increase of total energy by controlled drying process, as it is seen in figure 11, gives evidence that during this process no any cracks occurred.

Figure 11: The total energy registered during the controlled drying process.

The gentle increase of total energy, without rapid jumps as it is visible on the energy curve in figure 4, bespeaks that the material dried in a controlled way is slightly stressed. The AE signals, and in particular the energy transported by these signals prove that material is longer time under a state of stress. However, it is difficult to find a direct relationship between the magnitude of stresses and the emitted during drying energy.

The controlled drying process points out that in order to make the process economically efficient and save, as far as it concerns material fracture, the current analysis of AE signals have to be followed by suitable change of drying parameters in the course of drying. The controlled process with intermittent heating in a thermal drying is a one way to improve energy utilization and to enhance the quality of dried products. Efficient use of energy in drying process is of great economic benefit.

FINAL REMARKS

The main goal of this paper was to demonstrate the acoustic emission as a method useful for monitoring of the mechanical state arising in materials during drying. The studies carried out in this work on the pinewood show that the intensity of AE events depends on the drying conditions. The experiments suggest an immediate relation between the amount of emitted energy (less the number of AE events) and the destruction of the material. By convective drying, particularly when the drying proceeds in high temperature and small relative humidity of the drying medium, the thermodiffusional flux of moisture blockades the outflow of moisture due to diffusion, mainly at the boundary, and this causes strongly non-uniform distribution of the moisture content, what results in generation of stresses. The stresses are responsible for fracture of the dried material. In order to foreseen the fracture one ought to observe the number of AE events and, in particular, the amount of emitted energy. Making use of the AE method, we can control drying process and choose the drying conditions in such a way in order the drying induced stresses were possible small. In this paper the control was realized through changes in temperature program. When the weaker stresses are generated during drying a better quality dry product from the mechanical standpoint is obtained. Thus, the acoustic emission method may be proposed as a very useful method by construction of optimal drying processes.

ACKNOWLEDGEMENT

This work was carried out as a part of the research project DS 32/140/2004 sponsored by the Poznań University of Technology.


NOTE
This paper was first presented at the IDS-2004, Sao Paulo and pre-selected for Maderas.Ciencia y Tecnología journal.


LITERATURE
        [ Links ]

KAGAWA, Y.; NOGUCHI, M.; KATAGIRI, J. 1980. Detection of acoustic emission in the process of timber drying. Acoust. Lett. 3(8):150-153        [ Links ]

KITAYAMA, S.; NOGUCHI, M.; SATOYOSIHI, K. 1985. Monitoring of wood drying process by acoustic emission. Wood Ind. 40(10), 464-469        [ Links ]

KOWALSKI, S.J. 2003. Thermomechanics of Drying Processes. Springer-Verlag Berlin Heidelberg New York.        [ Links ]

KOWALSKI, S.J.; MOLINSKI, W.; MUSIELAK, G. 2004. The identification of fracture in dried wood based on theoretical modelling and acoustic emission. Wood Science and Technology 38:35-52         [ Links ]

NOGUCHI, M.; KAGAWA, Y.; KATAGIRI, J. 1980. Detection of acoustic emission during hardwood drying. Mokuzai Gakkaishi 26(9):637-638        [ Links ]

SKAAR, CH.; SIMPSON, W.T.; HONEYCUTT, R.M.1980. Use of acoustic emission to identify high levels of stress during oak lumber drying. For. Prod. J. 30(2):21-22[         [ Links ]STANDARDIZEDENDPARAG]

^rND^sCUNDERLIK^nI^rND^sMOLINSKI^nW^rND^sRACZKOWSKI^nJ^rND^sKAGAWA^nY^rND^sNOGUCHI^nM^rND^sKATAGIRI^nJ^rND^sKITAYAMA^nS^rND^sNOGUCHI^nM^rND^sSATOYOSIHI^nK^rND^sKOWALSKI^nS.J^rND^sMOLINSKI^nW^rND^sMUSIELAK^nG^rND^sNOGUCHI^nM^rND^sKAGAWA^nY^rND^sKATAGIRI^nJ^rND^sSKAAR^nCH^rND^sSIMPSON^nW.T^rND^sHONEYCUTT^nR.M^rND^sCUNDERLIK^nI^rND^sMOLINSKI^nW^rND^sRACZKOWSKI^nJ^rND^sKAGAWA^nY^rND^sNOGUCHI^nM^rND^sKATAGIRI^nJ^rND^sKITAYAMA^nS^rND^sNOGUCHI^nM^rND^sSATOYOSIHI^nK^rND^sKOWALSKI^nS.J^rND^sMOLINSKI^nW^rND^sMUSIELAK^nG^rND^sNOGUCHI^nM^rND^sKAGAWA^nY^rND^sKATAGIRI^nJ^rND^sSKAAR^nCH^rND^sSIMPSON^nW.T^rND^sHONEYCUTT^nR.M^rND^1A01^nBram J^sRamakers^rND^1A01^nRonny^sde Ridder^rND^1A01^nPiet J.A.M^sKerkhof^rND^1A01^nBram J^sRamakers^rND^1A01^nRonny^sde Ridder^rND^1A01^nPiet J.A.M^sKerkhof^rND^1A01^nBram J^sRamakers^rND^1A01^nRonny^sDe Ridder^rND^1A01^nPiet J. A. M^sKerkhof

Maderas. Ciencia y tecnología, 6(2): 145-153, 2004

ARTICULO

FLUIDIZATION BEHAVIOR OF WOOD/SAND MIXTURES

Bram J. Ramakers1, Ronny de Ridder1, Piet J.A.M. Kerkhof1

1Eindhoven University of Technology, Faculty of Chemical Engineering and Chemistry, Transport Phenomena Group, Laboratory of Chemical Reactor Engineering, P.O. Box 513, NL 5600 MB, Eindhoven, The Netherlands.


Corresponding author:
b.j.ramakers@tue.nl
; r.de.ridder@tue.nl ; p.j.a.m.kerkhof@tue.nl

Received: 05.10.2004. Accepted: 03.12.2004.

ABSTRACT

In conversion of biomass to secondary energy carriers, several routes are possible, such as gasification, combustion and pyrolysis. In many of these processes it is necessary or advantageous to dry the biomass before further processing. For wooden biomass, fluidized bed drying in superheated steam is a promising option. Given the difficulty to fluidize wood particles alone, it is very common to fluidize these kinds of particles with sand. This also gives better defined fluidization behavior. Especially when the wood particles come in various size and shape (i.e. from sawdust to chopped wood), this gives a more reliable scale-up. Also heat transfer to the wood particles may benefit from the use of sand. However, not much is known about fluidization behavior in pressurized steam of binary mixtures with large particle size ratio and large particle density ratio. Therefore minimum fluidization velocity and bed porosity of wood/sand mixtures in air have been experimentally determined and compared to correlations known from literature. The experimental values show a clear trend, but correlations from literature appear not to be very accurate. So more experiments have to be done to find a correlation that gives more accurate predictions in case of the specific particles used in this work. From segregation experiments could be found that, to keep the wood/sand bed well-mixed, finer sand (0.1-0.5 mm) with maximum 10 weight-% wood should be used, and the superficial gas velocity should be at least 3-4 times the minimum fluidization velocity.

Keywords: drying of wood, fluidization, binary mixture

INTRODUCTION

The desire for a sustainable society has led to research and development activities on the utilization of renewable energy sources. Biomass is considered to be such a resource: during the growth of plants and trees solar energy is stored as chemical energy, which can be released via direct or indirect combustion. In the overall process, CO2 is fixed during the production of biomass and released again during combustion. Within this relatively short cycle, no net addition of CO2 to the atmosphere takes place.

Conversion of biomass (after a preceding drying step) to intermediate energy carriers like gases or oils is probably much more attractive than direct combustion. Biomass as such is not very suitable for transport over long distances because it has a low bulk density, and liquefaction to an energy carrier such as oil, gas or electricity is an obvious solution. In biomass conversion there are several alternative routes, such as gasification, combustion, HTU-process, pyrolysis, etc.

In many of these processes it is necessary or advantageous to dry the biomass before further processing. From an energetic point of view such a drying step severely influences the economy of the whole plant. It is therefore necessary to optimize the drying step with respect to equipment, process conditions and pre-treatment; this optimization should be embedded in the optimal design and operation of the whole plant. In literature it is reported that drying with superheated steam may reduce both energy costs and the size of the drying equipment. For particulate biomass materials fluid bed steam drying is a promising option. The optimal design and operation rules for such a drying step can at present not be found from existing knowledge, but require more research.

As wood chips are rather difficult to fluidize, sand can be used as fluidizing medium. This also gives the benefit of a better described system in scale-up, when wooden particles of various sizes and shapes are used. Furthermore the sand will improve heat transport to the wooden particles in fluid bed drying. However, little is known about fluidization of binary mixtures with large size ratio and large density ratio. Therefore fluidization characteristics of various wood/sand mixtures are measured and fluidization behavior is observed, with cold air as fluidizing medium. These preliminary results will be used in designing a steam fluidized bed, in which fluidization characteristics of wood/sand mixtures in (superheated) steam will be measured, and in which the superheated steam fluidized bed drying of wood will be studied.

MATERIALS & METHODS

Theory

In the case of binary (polydisperse) systems, the minimum fluidization velocity umf (and fluidization“quality”) of large and/or coarse particles can be reduced by the addition of smaller particles. To estimate umf for a binary system, it is necessary to take into account the different characteristics of mixing and segregation of two kinds of particles having different properties. A lot of studies have been done about fluidization of binary systems with large particle size ratio, as well as about fluidization of binary systems with large particle density ratio. But little is known about fluidization of binary systems with both large size ratio and large density ratio. Most correlations to calculate umf are of the Ergun equation type (Kunii, and Levenspiel 1969; San José et al., 1995; Pell 1990; Reina et al., 2000; Gautier et al., 1999), and many of these require precise knowledge of the shape factor and bed porosity at umf. These two parameters are very difficult to measure experimentally and their values can have high errors that cannot be estimated, especially when irregular and coarse particles are handled (i.e. demolition wood or other biomass). Therefore the use of these parameters should be avoided in predicting umf for wood/sand mixtures. Several alternatives for predicting the minimum fluidization velocity of a binary mixture are listed below.

Goossens et al. (1971) proposed the following equations to calculate the density and particle size of a binary mixture, which can be used in Reynolds-Archimedes correlations:

(1)



(2)

The Wen and Yu (Wen and Yu, 1966; Noda et al., 1986) equation for a single component system is given by:

(3)

Chyang and Huang (San José et al., 1995) found the following relation for coarse particles:

(4)

Tannous et al. (San José et al., 1995) proposed the following equation for calculating the minimum fluidization velocity of coarse particles:

(5)

Chiba et al. (1979) derived an equation for umf of the mixture which depends on umf of the single components:

(6)

Experimental setup

The experimental setup (figure 1) consists of a glass pipe of 7.56 cm in diameter and a height of 120 cm. The distributor is a metal sintered plate of 5 mm thickness. The superficial velocity of cold air can be adjusted between 0 and 2.2 m/s. During fluidization pressure drop over the bed, absolute pressure before bed entrance and bed height can be measured.

Figure 1. Diagram of the fluidized bed setup; (1) glass fluid bed column, (2) flow meter, (3) pressure drop meter, (4) dust bag, (5) distributor, (6) pressure meter.

The wooden particles used in the fluidization experiments were beech wood cylinders of 9 mm length and 6 mm in diameter. Three different size distributions of sand were used as fluidizing agent. The properties of the particles are listed in table 1. The diameter is listed as that of a volume-equivalent sphere.

Table 1: Properties of the used particles.

Particle

Diameter [mm]

Density [kg/m3]

Wood

7.9

722

Sand (fine)

0.1-0.5

2629

Sand (medium)

0.4-0.6

2669

Sand (coarse)

0.8-1.2

2608

Prior to each experiment, the mixture was fluidized at high velocity (~1.5 m/s) for a few minutes. Then the airflow was suddenly stopped to obtain a well-mixed packed bed, with which all experiments started. Experiments are performed to find the minimum fluidization velocities and bed porosities of wood/sand mixtures at different ratios. Also segregation experiments are performed to examine bed mixing as a function of wood/sand ratio, type of sand and superficial gas velocity. These experiments started with a well-mixed bed, and after 1 minute fluidizing at a given airflow the airflow was cut. The fraction of wood on top of the wood/sand bed was measured and used as a criterion of the quality of fluidization and mixing of the bed. In all experiments total bed mass was kept constant.

RESULTS & DISCUSSION

Fluidization parameters

Figure 2 shows plots for calculating the minimum fluidization velocity and the bed porosity at minimum fluidization velocity. The fluidization parameters found for the pure components are listed in table 2. The minimum fluidization velocities, calculated with the modified Wen and Yu (1966) equation, are also added to Table 2. It can be seen that the experimentally determined umf for the sand fractions are nicely positioned in the middle of the window of the fractions, calculated with equation (3). This should be the case, assumed that the sand fractions have a Gaussian distribution.

Figure 2. Plots for calculating the minimum fluidization velocity and the bed porosity at minimum fluidization velocity; (a) Sand (medium), 0 w-% wood; (b) Sand (fine), 20 w-% wood.

 

Table 2: Hydrodynamic parameters of the pure components.

Particle

umf [m/s]

e mf [-]

Wen and Yu (1966)

Wood

1.30

0.425

1.30

Sand (fine)

0.13

0.494

0.01-0.20

Sand (medium)

0.22

0.457

0.13-0.27

Sand (coarse)

0.51

0.436

0.42-0.69

In Figure 3 the results of experiments with wood/sand mixtures are shown. From these figures can be seen that at a higher weight fraction of wood, the minimum fluidization velocity rises. This is due to the fact that the wood particles negatively influence the fluidization. The bed porosity, however, seems to minimize at 5 weight percent wood, and for fine sand even at 20 w-%. Mixtures in general have lower packed porosity than their pure component, with a minimum around 1:1 volume ratio (Chiba et al., 1979). With the particles used in this study that would be at 20 weight percent. This can be seen in Figure 3c. However, coarse and medium sand/wood mixtures have higher difference in umf in relation to the pure sand than in the case of fine sand/wood. Therefore their mixtures have a more expanded bed at umf, and the minimum in e mf should shift to lower weight fractions of wood (Figures 3a and 3b).

Figure 3. Minimum fluidization velocity and bed porosity at umf as function of the weight fraction of wood; (a) coarse sand/wood; (b) medium sand/wood; (c) fine sand/wood.

The values for umf determined in this work are compared to values calculated from literature correlations, stated in equations (1) to (6) (Figure 4). It can be clearly seen that in most cases there is an over or under prediction of umf. Wen and Yu, equation (3), appears to give the best prediction for all three types of sand. However, to get a more accurate prediction for umf, more experiments should be done to derive a correlation for umf that could be used for the specific particles used in this work. This is due to the great influence of shape, sizes and size ratio, and densities and density ratio of the particles on the fluidization behavior of the mixture.

Table 3: Uncertainties in measured experimental data.

Quantity

Estimated error

Pressure difference, D P

± 5 Pa

Minimum fluidization velocity, umf

± 0.02 m/s (Fine); ± 0.03 m/s (Medium);

± 0.05 m/s (Coarse)

Bed porosity at umf, e

± 0.01 (Fine); ± 0.015 (Medium); ± 0.025 (Coarse)

Segregation experiments

The results of the segregation experiments are shown in Figure 5. From these experiments it can be seen that less wood in the system, keeps the bed mixed at lower superficial gas velocities. Also the finer sand seems a better fluidizing agent for the wood particles used. For the future drying experiments there has to be a reasonable amount of wood in the bed. From this work, 10 weight percent wood in fine sand seems to give best prospect. Nevertheless, the superficial gas velocity has to be 3-4 times higher than umf in order to keep the bed well-mixed.

Figure 4. Minimum fluidization velocity found in this work compared to literature, equations (1) to (6).


CONCLUSIONS

The measurements of fluidization behavior of binary mixtures with large size ratio and large density ratio show clear trends in umf as well as in e mf. However, it appears that using correlations from literature for calculating umf, doesn’t predict the umf very well for the specific mixtures used in this work. More experiments should be done to derive a better correlation. Especially when in future work superheated steam will be used to fluidize wood/sand mixtures, the fluidization behavior of these specific mixtures in steam should be experimentally determined.

The segregation experiments show that in fluid bed drying of the wood particles used, fine sand is best used as fluidizing agent. Also the superficial gas velocity should be at least 3-4 times umf to have a well-mixed bed. The weight fraction of wood should then be 0.10 at most to keep the bed from segregating.

 

ACKNOWLEDGMENT

The authors would like to kindly thank the Council for Chemical Sciences of the Netherlands Organization for Scientific Research (CW-NWO, projectnr. 99602), Shell Global Solutions International B.V. and the Netherlands Organization for Energy and Environment (Novem) for funding this research.


NOTE

♣This paper was first presented at the IDS-2004, Sao Paulo and preselected for MADERAS:Ciencia y Tecnología journal.

LITERATURE


CHIBA, S.; CHIBA, T.; NIENOW, A.W.; KOBAYASHI, H. 1979. The minimum fluidization velocity, bed expansion and pressure-drop profile of binary particle mixtures. Powder Technology 22:255-269        [ Links ]

GAUTHIER, D.; ZERGUERRAS, S.; FLAMANT, G. 1999. Influence of the particle distribution of powders on the velocities of minimum and complete fluidization. Chemical Engineering Journal 74:181-196        [ Links ]

GOOSSENS, W.R.A.; DUMONT, G.L.; SPAEPEN, G.L. 1971. Chemical Engineering Progress Symposium Series 67 (116):38-45        [ Links ]

KUNII, D.; LEVENSPIEL, O. 1969. Fluidization Engineering. Wiley, New York        [ Links ]

NODA, K.; UCHIDA, S.; MAKINO, T.; KAMO, H. 1986. Minimum fluidization velocity of binary mixtures of particles with large size ratio. Powder Technology 46:149-154        [ Links ]

PELL, M. 1990. Gas fluidization, Handbook of Powder Technology. Vol. 8, Elsevier Science Publishers B.V., The Netherlands        [ Links ]

REINA, J.; VELO, E.; PUIGJANER, L. 2000. Predicting the minimum fluidization velocity of polydisperse mixtures of scrap-wood particles. Powder Technology 111:245-251        [ Links ]

SAN JOSÉ, M.J.; OLAZAR, M.; BENITO, P.L.; BOLBAO, J. 1995. Hydrodynamics and expansion of fluidized beds of coarse particles. Transactions of the Institution of Chemical Engineers Vol. 73, part A:473-479        [ Links ]

WEN, C.J.; YU, Y.H. 1966. A generalized method for predicting the minimum fluidization velocity. AIChE Journal 12(3):610-612         [ Links ]

 

NOTATION

Ar

Archimedes number

-

d

Particle diameter

 

m

Re

Reynolds number

-

u

Superficial gas velocity

 

m/s

ε

Bed porosity

 

-

μ

dynamic viscosity

 

kg/ms

ρ 

fluid density

 

kg/m³

v

Weight fraction

 

-

     

Subscripts

 

Superscripts

 

mf

Minimum fluidization velocity

F

Small particle

m

mixture

P

Large particle

       

 

^rND^sCHIBA^nS^rND^sCHIBA^nT^rND^sNIENOW^nA.W^rND^sKOBAYASHI^nH^rND^sGAUTHIER^nD^rND^sZERGUERRAS^nS^rND^sFLAMANT^nG^rND^sGOOSSENS^nW.R.A^rND^sDUMONT^nG.L^rND^sSPAEPEN^nG.L^rND^sNODA^nK^rND^sUCHIDA^nS^rND^sMAKINO^nT^rND^sKAMO^nH^rND^sREINA^nJ^rND^sVELO^nE^rND^sPUIGJANER^nL^rND^sSAN JOSÉ^nM.J^rND^sOLAZAR^nM^rND^sBENITO^nP.L^rND^sBOLBAO^nJ^rND^sWEN^nC.J^rND^sYU^nY.H^rND^sCHIBA^nS^rND^sCHIBA^nT^rND^sNIENOW^nA.W^rND^sKOBAYASHI^nH^rND^sGAUTHIER^nD^rND^sZERGUERRAS^nS^rND^sFLAMANT^nG^rND^sGOOSSENS^nW.R.A^rND^sDUMONT^nG.L^rND^sSPAEPEN^nG.L^rND^sNODA^nK^rND^sUCHIDA^nS^rND^sMAKINO^nT^rND^sKAMO^nH^rND^sREINA^nJ^rND^sVELO^nE^rND^sPUIGJANER^nL^rND^sSAN JOSÉ^nM.J^rND^sOLAZAR^nM^rND^sBENITO^nP.L^rND^sBOLBAO^nJ^rND^sWEN^nC.J^rND^sYU^nY.H^rORG^1A01^nMaria Aparecida^sSilva^rORG^1A01^nMaria Aparecida^sSilva^rORG^1A01^nMaria Aparecida^sSilva

Maderas. Ciencia y tecnología, 6(2): 155-158, 2004

COMENTARIO

14th International Drying Symposium (IDS 2004) Conference Report

Maria Aparecida Silva1
1IDS 2004 Chairwoman. Sao Paulo. Brasil

Corresponding author: ids2004@feq.unicamp.br


Over almost 30 years of existence of the International Drying Symposia (IDS) the number of Brazilian participants increased continuously. The best way to reflect the increasing position of Brazilian drying research was to host the 14th International Drying Symposium (IDS 2004) in São Paulo City, Brazil during August 22-25, 2004; this was the first time the symposium was hosted by a Latin American country.

The symposium welcomed 221 participants from 37 countries from the five continents. As in all previous IDS’s, since its first edition in 1978, a truly international forum was guaranteed and interesting and instigating discussions on drying were carried out. IDS 2004 was a little bit smaller in number of participants than the last IDS’s. Nevertheless, the participants were very active: more than 50 people in each one of the three parallel oral sessions. In the poster sessions more than 200 participants discussed the papers for more than 2 hours. And even the closure session had more than 150 participants.

The 3-volume hardcover set of proceedings (also in CD-ROM) is constituted by a total of 279 papers (1 plenary lecture, 6 invited keynote lectures and 272 contributed papers) written by 648 authors from 44 countries. There were 26 joint papers with authors from different countries. Figure 1 shows the distribution of authors of IDS 2004 papers by their country of origin. France and Japan, as usual, appeared as the biggest contributors after the host.

Figure 1: Country of origin of IDS 2004 authors

Figure 2 presents the participants of IDS 2004 also by their country of origin. Surprisingly, Sweden and Poland contributed with the biggest number of participants just after the host. As expected, the host was the greatest but never in previous IDS’s the host presented such a great contribution as in IDS 2004, around 40% of the authors, papers and participants came from Brazil. It is important to remark that Brazil has been always the second/third contributor of IDS’s since IDS’96.

Figure 2: Country of origin of IDS 2004 participants

 

The participation of people from industry was again not so big, only 13% of the participants came from industry. In fact, only two IDS’s had high industry participation: IDS’78 and IDS 2000.

IDS 2004 was the first IDS that all information, abstract submission, manuscript submission, reviewing process and registration were electronically processed. The system worked quite well since more than 400 persons accessed it and less than 2% found problems of incompatibility.

A total of 430 abstracts were submitted to the Conference Secretariat while 316 manuscripts were received. After the reviewing process, 32 were rejected and 284 returned to the authors for corrections. Finally, 276 received the final acceptance, however, 4 were excluded from the program because none of their authors sent the registration form although several messages and faxes were sent to them. So, 272 papers were published in the proceedings and kept in the IDS 2004 program. A total of 25 reviewers work very hard in the reviewing process, being 15 from the Scientific Committee (all members of the Advisory Panel of IDS’s) and 10 from the Organizing Committee. Each manuscript was reviewed by 2 of them.

Seventy-eight papers were distributed in 14 oral presentation sessions. Eight presentations were scheduled for the special session of softwares. A hundred and ninety four papers were scheduled for 2 sessions of poster presentations. An impressive characteristic of IDS 2004 was practically the absence of no-shows. 94% of the papers scheduled for oral session were presented against 95% of the posters. By chance the papers not presented in the oral session were scheduled to the end of the sessions, so their absence didn’t hold the participants from moving from one session to another.

Figure 3 shows the distribution of the IDS 2004 papers according to their themes. As in previous IDS’s, the predominance of Drying of Food and Agricultural Materials along with Fundamental/modeling& simulation was observed.

Figure 3: Themes of the IDS 2004 papers

The distribution of the authors, papers and participants of IDS 2004 by their continent of origin can be observed in Figure 4. It is clear that the interest in drying research is still high in Asia and Europe. Of course, South America appears in the first place due to the huge participation of the host. Although the distance could have helped the contribution of North America, this didn’t happen, reinforcing the idea of declining interest in drying research by North Americans.

Figure 4: Distribution of authors, papers and participants of IDS 2004 by their continent of origin

Having in mind to offer to IDS 2004 attendants different insights of the future of the drying development, it was planned the Open Forum entitled "New trends in drying research and technology", and it happened a very interesting and productive discussion between the audience and the speakers.

It was a great pleasure to receive people from different countries in IDS 2004 sharing with them a typical Brazilian ambiance. We are sure the symposium promoted many new contacts and permitted to renew old ones.

IDS 2004 website presents information about the symposium including the program and the proceedings. http://www.feq.unicamp.br/~ids2004.

The next International Drying Symposium, IDS 2006, will be held in Budapest, Hungary, 20-23 August 2006. http://fft.gau.hu/events/ids2006.html




NOTE
♣ 14th International Drying Symposium (IDS 2004) São Paulo City, Brazil August 22-25, 2004.

 

HP-2

2.52

2.36

1.54

Average SEC*** (kWh/kgwater )

HP-1

0.42

0.47

0.68

HP-2

0.40

0.42

0.64

HP – Heat Pump; COP* – Coefficient of Performance; SMER** – Specific Moisture Extraction Rate (based on compressor and blower energy consumption); SEC*** – Specific Energy Consumption (based on compressor and blower energy consumption).

Operating Lessons Learned

Spread over several months, the first development step of high-temperature heat pump prototypes was aimed at ensuring a maximum operational stability of the thermodynamic parameters, establishing the reliability of the most sensible components (compressors, blowers, safety valves), checking the refrigerant/oil blend’s behavior and optimizing the critical control sequences of the system. The experimental dehumidifier dryer operated in “extreme” temperature, humidity and corrosion conditions, and demonstrated specific feedback phenomena that do not occur in traditional air-forced/heated lumber dryers. Because some of the by-products that could be produced by chemical interactions between the lubricant and working fluid may be acidic and lead to accelerating the corrosion of the system components, periodical controls aimed to determine the chemical behavior of the mixture. After about 3,250 hours of operation, the refrigerant proved to be thermally stable and chemically inert at the highest temperatures occurring in the system and a first oil chemical analysis proved that there were no problems with the oil breaking down or failing. The oil still showed adequate viscosity and chemical stability as well as a good miscibility with the refrigerant. The initial designed capacity of the heat pumps proved to be too high and consequently, both compressors were slowed down by about 25% which finally resulted in a more adequate capacity, reduced head pressures and improved efficiency. Because the original employed expansion valves poorly controlled the refrigerant flowing, did not open fast enough and manufacturer leakage faults were detected, they were later replaced by new generation devices. A crack in the casting of an original pressure relief valve also indicated a manufacturer’s defect, and finally both these components have been replaced to 20% higher pressure-limit valves. Other issues included the fact that the kiln was initially poorly insulated and leaky, and some system’s components were prematurely corroding.

CONCLUSIONS

As a clean energy technology compared with traditional heat-and-vent dryers, high-temperature heat pump dehumidifiers offer interesting benefits for drying resinous timber. This paper presents the preliminary results of the development and field testing of two prototypes accentuating their thermodynamic parameters, preliminary energy performances and first operating lessons learned. The average measured specific moisture extraction rate of the heat pumps was 2.35 kgwater/kWh (white spruce) and 1.5 kgwater/kWh (balsam fir), while the average coefficients of performance generally varied from 3.0 to maximum of 4.6. The cycle’s duration ranged from 2.5 days (white spruce) to 6.3 days (balsam fir) including the initial preheating steps. The refrigerant/oil mixture behaved well during more than 3250 hours of preliminary tests, proving good compatibility and chemical stability at condensing temperatures below 110°C (230°F). Better insulated and well maintained dryers are necessary to obtain drying temperatures higher than 100°C (212°F) as well as reducing the drying duration of resinous species by up to 25% and the total energy consumption by up to 50%. The current goals of the study include using more corrosive resistant components, variable speed central fan, further optimizing the drying schedules and general dryer operation and maintenance. Finally, it is expected to help local Canadian equipment suppliers to promote research and development of the technology and develop an appropriate market strategy. Specifications of high-temperature heat pump dehumidifier kiln energy use and a best-practice guideline must also be produced.

ACKNOWLEDGEMENTS

The author gratefully acknowledges “Hydro-Quebec’s Energy End-Use Service” for their indefectible support in this R&D project and our technician, Mr. Marcel Déry, who installed the monitoring system, helped in the long-term survey and contributed to the enhancement of the system’s electronic controls. Finally, the availability of the heat pumps, their field development and comprehensive testing were made possible thanks to a Canadian sawmill and to a North American heat pump manufacturer.

NOTE

♣ This paper was first presented at the IDS-2004, Sao Paulo and pre-selected for MADERAS:Ciencia y Tecnología journal.

LITERATURE

CANADA STATISTICS AND INDUSTRY MINISTER. 2002. Public Documents        [ Links ]

CECH, M.J; PFAFF, F. 2000. Operator Wood Drier Handbook for East of Canada. Edited by Forintek Corp., Canada’s Eastern Forester Products Laboratory        [ Links ]

KASACHKI, G.S.; GAGE, C.L; HENDRIKS, R.V. 1994. Investigation of HFC-236ea and HFC-236fa as CFC-114 Replacements in High-Temperature Heat Pumps. CFC’s: The Day After, Padua, Italy[         [ Links ]STANDARDIZEDENDPARAG]

^rND^1A01^nVasile^sMinea^rND^1A01^nVasile^sMinea^rND^1A01^nVasile^sMinea

Maderas. Ciencia y tecnología, 6(2): 123-132, 2004

ARTICULO



HEAT PUMPS FOR WOOD DRYING – NEW DEVELOPMENTS AND PRELIMINARY RESULTS

1Vasile Minea
1Ph.D. Institut de Recherche d’Hydro-Québec, Laboratoire des Technologies de l’Énergie (LTE), 600, de la Montagne, Shawinigan, G9N 7N5, Canada.

Corresponding author:
minea.vasile@lte.ireq.ca

Received: 05.10.2004. Accepted : 22.11.2004.


ABSTRACT

This paper succinctly presents new developments, preliminary statements and a number of energy results in the area of high-temperature heat pump technology for wood drying in a Canadian economic environment. A hybrid (electricity/fossil), high-temperature technology has been investigated and then field tested over the last two years. Several technical developments were achieved at the level of fluid selection, refrigerant flow control and system stability, variable dehumidifying capacity and appropriate drying schedules. The present study demonstrates that the thermodynamic efficiency and specific energy performance of the developed high-temperature drying heat pumps have generally reached the initial designed targets. Refinements of the integrated control methods involving variable speed and electronic devices are currently being undertaken in order to avoid undesired operating conditions that could cause mechanical failures or inefficient dehumidifying processes. The current research program aims at diversifying the applicable thermodynamic cycles, testing new environmentally friendly refrigerants and advanced components, and developing more advanced drying control strategies.

Keywords: Wood Drying

INTRODUCTION

The drying of resinous lumber, a process that is highly favorable to high temperatures, is essential to prevent the warping and cracking of the wood. In 2002, Canadian sawmills delivered 72 million such products, approximately 10% of which were produced in the province of Quebec (East of Canada) and where only 2% were dried by heat pumps and the rest through other technologies such as direct fire and the use of bark-, natural gas- or oil-burned boilers [Canada Statistics 2002]. In the 1970s and 1980s, the kiln drying industry promoted dehumidifier concepts, but the performance of such systems was often disappointing due to air flow and control problems, inadequate dehumidifying capacity and inappropriate kiln structures. The reliability of the systems was often low and equipment suppliers did not provide enough information on the actual performance of their systems. However, the best solution that would allow heat pumps to be effective as a means of drying wood would be to use them in combination with a traditional energy source at high temperatures to obtain similar drying speeds and energy savings. The technology of high-temperature heat pumps is not yet available, all the more so since early 90s environmental issues imposed the replacement of traditional CFCs used as high-temperature refrigerants. The merits of heat pump dryers however include lower energy consumed for each unit of water removed, accurate control of drying conditions, and enhanced product quality. Their limitations generally concern the need for regular maintenance, the risk of refrigerant leaks and higher initial capital costs compared to conventional dryers. Recently, more environment-friendly refrigerants have been developed and, two years ago, a North American heat pump manufacturer and a Canadian lumber producer decided to put their field experience and know-how to use in order to develop a high-temperature heat pump dryer application and to experiment their prototypes on an industrial scale, even if the processes involved in wood drying are highly non-linear and, consequently, the scale-up of dryers is generally difficult. Hydro-Quebec’s Research Institute has actively contributed to the development of the first two industrial prototypes and to a field testing program aimed at providing customers with quantitative information on the reliability and efficiency of such systems.

DRYING SYSTEM CONFIGURATION

The sawmill facility chosen as the experimental site has been equipped since 1998 with two air forced, traditional -wood dryers made of insulated panels, each including 1,500-kW steam heating coils (Figure 1). An oil-burned boiler of 4,900-kW output capacity (82%) supplies both dryers with high-pressure saturated steam for heating and spraying. One of these dryers, where a maximum of 96 to 100 bundles enter on two train rails, has been converted into a hybrid dryer equipped with two high-temperature heat pumps and a back-up steam heating system.

Figure 1 –View of the Experimental Wood Dryer (on left)

Charged at full capacity during the cold Canadian climate, the experimental dryer is theoretically able to generate drying temperatures of up to 116°C (240°F) in a lapse of 6 to 8 hours. However, heat losses and leakage technically lead to maximum dry temperatures of 82.2°C to 93.3°C (180°F to 200°F). A 56-kW low static pressure, longitudinal, six-blade fan with an outdoor motor, placed above a false ceiling ensured forced circulation of the air at 1.5 – 2.0 m/s (300 – 400 fpm) at the stacks of wood outlet. Mural deflectors and inversion of the rotation of the central fan at every 3 hours at the beginning and at every 2 hours at the end of the drying cycles contribute to obtaining uniform ventilation. Nine of the twelve existing air vents, placed in two rows on each side of the longitudinal fan, are kept closed in the hybrid dryer. To avoid air implosion risks, the three operational air vents open solely when the central fan changes its rotation direction and also when the actual dry temperature exceeds the set point. The high-temperature heat pumps, each equipped with a variable speed blower, are linked to the hybrid dryer (Figure 2). Compressors, evaporators and their electric/electronic controls are placed inside the adjacent mechanical room, while the condensers, as one design originality, are installed inside the kiln. Designed for industrial processes, the open-, belt-driven compressors are provided with oil pumps, external pressure relief valves and crankcase heaters. The used refrigerant, a non-toxic and non-flammable fluid, readily available in Canada and relatively inexpensive compared to conventional refrigerants, has a relatively high critical temperature compared to the highest process temperature and a normal boiling point les than the lowest temperature likely to occur in the system. Moreover, the saturation vapour pressure at highest design temperature is not so high as to impose design limitations on the system. Previous studies have shown that it has cooling capacities of up to 20% higher than the best older high-temperature refrigerants [Kasachki 1994]. Multiple, parallel installed expansion valves are incorporated into the microprocessor-based temperature/process controllers that display the set point and the actual process temperatures.

Figure 2 – General Layout of the Experimental Drying Plant
C – Compressor. EV – Evaporator. CD – Condenser. F – Fan

Drying Schedules

Humidity exists inside the wood boards as “free” (liquid or vapor) and “linked” (hygroscopic) water and it is practically admitted that the fiber saturation point is 30% of the dry basis moisture content. Displacement of the humidity through the wood is generally driven by affinity and capillarity (adhesion/cohesion) forces, vapor and moisture content gradients and diffusion, while the drying velocity is firstly governed by the ambient air capacity to absorb humidity and widely depends on the temperature and relative air and wood dryness. Three of several preliminary tests performed are presented here, respectively with white spruce Picea glauca (#70 and #88) and balsam fir Abies balsamea (#176) (see Table 1). All these batches were submitted to preheating periods at a maximum 87.7°C (190°F) dry temperature before each first step of the drying cycle, generally for a period of 6 to 8 hours in order to destroy the micro-organisms responsible for discoloring the sapwood. Presently, all preheating steps are performed at 93.3°C (200°F) dry temperatures. When the heat pumps started up (step 1), the moisture content first decreased linearly with time, a process followed by a non-linear decrease until the wood boards reached their equilibrium state and the drying cycle then shut down at the end of 5th or 6th step. The drying conditions of each step were established based upon moisture content, type of wood species, dimensions and quality of the wood, in conformity with an index established for Eastern Canada wood drying programs [Cech & Pfaff 2000]. For white spruce, which is normally easy to artificially dry, at initial moisture content of between 40 and 30%, the setting dry bulb temperature normally was 82.2 to 85°C (180 to 185°F) and the wet bulb temperature, 62.7°C (145°F). At a moisture content of less than 30%, the dry bulb temperature was generally 79.4°C (175°F) and the wet bulb temperature, 62.7°C (145°F). However, with balsam fir, which is harder to dry, when the initial moisture content was above 35%, the dry bulb temperature remained at 82.2°C (180°F) and the wet bulb at 79.4°C (175°F). Finally, for moisture contents lower than 25%, the setting dry bulb temperature attained 93.3°C (200°F) whereas the wet bulb temperature was 71.1°C (160°F). Changes in dry and wet temperatures settings during the tests were done on predetermined time-based schedules. For white spruce, steps 1 to 3 generally lasted 10 hours, while step 4 held out 20 hours, and step 5, 10 or even 20 hours depending upon the wood’s actual moisture content. In the case of balsam fir, the first five drying steps each lasted 30 hours, while the 6th step lasted up to 15 hours. The main idea was to not exceed the average duration of traditional drying batches for the same species of dried wood. Finally, when the indoor dry bulb temperature was lower than the set point value, the steam valve opened gradually from 5% to100% upon a time-based schedule to fully recover the set point. However, some work should still be focused on improving the steam valves and air vent actual operation.

Preliminary results

Thermodynamic Parameters

Heat pumps have normally not cycled at the start of preliminary drying processes, but they sometimes did that at the end as a result of the reduced ability of the wood to give up water at low moisture contents and the scheduled drop of the wet bulb temperature. The actual control strategy effectively allowed compressors to shut-down when the actual drier wet temperature has reached the set-point (Figure 3). After a given time delay, they were allowed to restart if the wet bulb temperature exceeded the presetting set point and only when a minimum presetting suction pressure was detected.

Figure 3 – Profiles of Dryer Wet Bulb Setting and Actual Temperatures (Example: Batch #88)

The duration of drying batch#88, graphically presented here as a typical example, was 61.3 hours, and the respective graphs represent data when the compressor was in operation. They do not include the approximately 6-hour timber preheating step, which increased the dryer dry bulb temperature to 87.7°C (190°F). Within the reduced capacities conditions, the compressors ran with shaft electrical powers varying between 60 kW and 65 kW, and average compression ratios of 5 to 6. Stable suction and discharge pressure as well as an average condenser sub-cooling of 8°C (14.4°F) were ensured. Typical condensing temperatures varied around 100–105°C (212–220°F), about 20°C (36°F) higher than the kiln dry bulb temperature, and the evaporating temperature was in the range of 41.1 - 45.5°C (106 -114°F). The average relative humidity of the air entering the evaporators largely varied because of periodical changes in the rotation direction of the central fan, and has also continuously decreased in time. However, the relative humidity leaving the evaporators was almost constant at around 74% to 88%, except at the end of the cycle when it dropped to 70% (Figure 4). All measurements shown are original data scanned at 15-second intervals and saved at 2-minute intervals. The preheating step, which is necessary to prevent the discoloration of sapwood, allowed increase the kiln’s absolute humidity up to 0.35 kg/kg before the dehumidifying process start-up (Figure 5). The maximum gradient of the absolute humidity across the heat pump’s evaporators, varied from 0.214 kg/kg, immediately after starting the compressors, to about 0.039 kg/kg at the end of the showed drying cycle. The pick drying rate and efficiency of the heat pump were significantly higher than the average values, due to a higher wood moisture production rate at the beginning of the cycle. As in previous graphs, all variations of the shown parameters are due to the changes in the fan’s direction of rotation that periodically modified the air flow pattern and humidity distribution inside the kiln.

Figure 4 – Air’s Relative Humidity Entering and Leaving the Evaporator
(Example: Heat Pump-1; Batch #88)

Figure 6 represents three typical dehumidifying processes (batch#88) based on the average measured parameters respectively at the beginning, at the end and as cycle’s averaged values. The gradient of the air’s mass enthalpy of the average drying cycle was about 184 kJ/kg, while the dry bulb temperature varied from 82.2°C (180°F) (inlet) to 54.4°C (130°F) (outlet), and the corresponding wet bulb temperature, from 65.8°C (140°F) to 48.8°C (120°F). Tests have yet to measure the actual oil consumption for preconditioning and back-up heating because the hybrid dryer was always in operation at the same time as the adjacent conventional dryer.

Figure 5 – Air’s Absolute Humidity Entering and Leaving the Evaporator
(Example: Heat Pump -1; Batch #88)

Energy Performances

Table 1 provides the average energy performances of both high-temperature heat pumps for three typical dehumidification cycles. The moisture content of the timber prior to drying was typically in the range of 35 - 45% indicating that it had been air dried for a few days in the facility yard prior entering to kiln drying. The heat pump’s average coefficient of performance (COP), defined as useful thermal power output (W) divided by electrical power input (W), varied from 3 to 4.6. The total average kiln rates of 313 kg water/hour (batch # 70), 263.2 kg water/hour (batch #88) and 178.8 kg water/hour (batch #176) do not include the venting losses that contributed on average approximately 90 kg water/hour (estimated value), but contained 5% of the estimated condensed water losses. The Specific Moisture Extraction Rate (SMER), defined as the amount of water extracted by the heat pump (kg) and the total energy input (compressor and blower) expressed in kWh, and that generally depends on the maximum air temperature, humidity, evaporator and condenser operating temperatures and the overall efficiency of the heat pump cycle, varied from 1.46 kgwater/kWh (batch #176) to 2.52 kgwater/kWh (batch#70), excluding preheating energy consumption. It did not include an allowance for the energy consumption of the kiln’s central fan or the venting moisture losses.

Figure 6 – Air Dehumidifying Processes (Example: Heat Pump-1; Batch #88)
EE – Entering Evaporator; LE – Leaving Evaporator; START – at the beginning of the cycle; AVRG – average drying cycle; END – at the end of the cycle

Another parameter, known as the Specific Energy Consumption (SEC), varied from 0.4 to 0.68 . The heat pump’s consumption (compressor and blower) represented about 72% of the total electrical energy consumption, the balance being for the dryer’s central fan (28%). As noted, at this time, any pertinent oil consumption measurement was done because the experimented hybrid kiln was always used in tandem with the neighboring conventional dryer, but instrumentation had already been installed on the oil supply line. The drying time to obtain white spruce with an approximate final moisture content of 17% to 19% averaged 2.5 days, while for balsam fir it averaged 6.3 days. Despite its longer duration, the drying of the balsam fir was less concerned with drying speed than that of the white spruce, the main operating focus values of which were to produce a high quality product. Last year, without heat pump dryers, the facility’s specific cost for an annual production of about 39,600 of dried lumber was 14.75 US$/ including all regular expenses (kiln operation, electrical and fossil energy consumption, equipment depreciation, insurance, etc.), of which energy only represented 6.86 US$/ . The facility’s objective is to reduce the energy-specific cost by at least 40%, and this also represents a goal of the current development steps related to high-temperature heat pumps.

Table 1 – Energy Performances of Three Preliminary Drying Cycles

Batch Number

-

#70

#88

#176

Parameter

Unit

-

-

-

Timber

-

White Spruce

White Spruce

Balsam Fir

Board Dimensions (cm x cm x m)

-

5.1x10.1x2.74

5.1x7.62x3

5.1x7.62x3

Duration of the Drying Cycle (hours) (Except preheating period)

HP-1

61.00

61.3

151.4

HP-2

61.00

61.3

151.4

Average Compressor Power (kW)

HP-1

65.12

63.36

61.0

HP-2

62.78

58.50

57.14

Compressor Energy Consumption (kWh)

HP-1

3,972

3,884

9,235.4

HP-2

3,830

3,586

8,651.0

Blower Energy Consumption (kWh)

HP-1

13.42

16.6

28.7

HP-2

14.03

39.5

107.5

Condensed Water Extracted (Liters)

HP-1

9,454

8,263

13,550

HP-2

9,655

8,478

13,531

Final Batch Average Moisture Content (%)

-

17.2

20.6

20.7

Average COP* (-)

HP-1

4.23

4.6

3.46

HP-2

3.70

4.07

3.00

Average SMER** (kgwater/kWh)

HP-1

2.38

2.13

1.46

HP-2

2.52

2.36

1.54

Average SEC*** (kWh/kgwater )

HP-1

0.42

0.47

0.68

HP-2

0.40

0.42

0.64

HP – Heat Pump; COP* – Coefficient of Performance; SMER** – Specific Moisture Extraction Rate (based on compressor and blower energy consumption); SEC*** – Specific Energy Consumption (based on compressor and blower energy consumption).

Operating Lessons Learned

Spread over several months, the first development step of high-temperature heat pump prototypes was aimed at ensuring a maximum operational stability of the thermodynamic parameters, establishing the reliability of the most sensible components (compressors, blowers, safety valves), checking the refrigerant/oil blend’s behavior and optimizing the critical control sequences of the system. The experimental dehumidifier dryer operated in “extreme” temperature, humidity and corrosion conditions, and demonstrated specific feedback phenomena that do not occur in traditional air-forced/heated lumber dryers. Because some of the by-products that could be produced by chemical interactions between the lubricant and working fluid may be acidic and lead to accelerating the corrosion of the system components, periodical controls aimed to determine the chemical behavior of the mixture. After about 3,250 hours of operation, the refrigerant proved to be thermally stable and chemically inert at the highest temperatures occurring in the system and a first oil chemical analysis proved that there were no problems with the oil breaking down or failing. The oil still showed adequate viscosity and chemical stability as well as a good miscibility with the refrigerant. The initial designed capacity of the heat pumps proved to be too high and consequently, both compressors were slowed down by about 25% which finally resulted in a more adequate capacity, reduced head pressures and improved efficiency. Because the original employed expansion valves poorly controlled the refrigerant flowing, did not open fast enough and manufacturer leakage faults were detected, they were later replaced by new generation devices. A crack in the casting of an original pressure relief valve also indicated a manufacturer’s defect, and finally both these components have been replaced to 20% higher pressure-limit valves. Other issues included the fact that the kiln was initially poorly insulated and leaky, and some system’s components were prematurely corroding.

CONCLUSIONS

As a clean energy technology compared with traditional heat-and-vent dryers, high-temperature heat pump dehumidifiers offer interesting benefits for drying resinous timber. This paper presents the preliminary results of the development and field testing of two prototypes accentuating their thermodynamic parameters, preliminary energy performances and first operating lessons learned. The average measured specific moisture extraction rate of the heat pumps was 2.35 kgwater/kWh (white spruce) and 1.5 kgwater/kWh (balsam fir), while the average coefficients of performance generally varied from 3.0 to maximum of 4.6. The cycle’s duration ranged from 2.5 days (white spruce) to 6.3 days (balsam fir) including the initial preheating steps. The refrigerant/oil mixture behaved well during more than 3250 hours of preliminary tests, proving good compatibility and chemical stability at condensing temperatures below 110°C (230°F). Better insulated and well maintained dryers are necessary to obtain drying temperatures higher than 100°C (212°F) as well as reducing the drying duration of resinous species by up to 25% and the total energy consumption by up to 50%. The current goals of the study include using more corrosive resistant components, variable speed central fan, further optimizing the drying schedules and general dryer operation and maintenance. Finally, it is expected to help local Canadian equipment suppliers to promote research and development of the technology and develop an appropriate market strategy. Specifications of high-temperature heat pump dehumidifier kiln energy use and a best-practice guideline must also be produced.

ACKNOWLEDGEMENTS

The author gratefully acknowledges “Hydro-Quebec’s Energy End-Use Service” for their indefectible support in this R&D project and our technician, Mr. Marcel Déry, who installed the monitoring system, helped in the long-term survey and contributed to the enhancement of the system’s electronic controls. Finally, the availability of the heat pumps, their field development and comprehensive testing were made possible thanks to a Canadian sawmill and to a North American heat pump manufacturer.

NOTE

♣ This paper was first presented at the IDS-2004, Sao Paulo and pre-selected for MADERAS:Ciencia y Tecnología journal.

LITERATURE

CANADA STATISTICS AND INDUSTRY MINISTER. 2002. Public Documents        [ Links ]

CECH, M.J; PFAFF, F. 2000. Operator Wood Drier Handbook for East of Canada. Edited by Forintek Corp., Canada’s Eastern Forester Products Laboratory        [ Links ]

KASACHKI, G.S.; GAGE, C.L; HENDRIKS, R.V. 1994. Investigation of HFC-236ea and HFC-236fa as CFC-114 Replacements in High-Temperature Heat Pumps. CFC’s: The Day After, Padua, Italy[         [ Links ]STANDARDIZEDENDPARAG]

Creative Commons License Todo o conteúdo deste periódico, exceto onde está identificado, está licenciado sob uma Licença Creative Commons