Modeling of thermal energy demand in MDF production.
A computer model has been developed to quantify the thermal energy flow in the production process of medium density fiberboard (MDF). The processing elements that consume thermal energy are grouped into three primary unit operations: chip preheating and refining, fiber drying, and mat hot pressing. With inputs of MDF annual production, plant operation hours, product grade, and fiber drying method, the model is able to predict the thermal energy demand and distribution, energy quality, and self-sufficiency level with using the wood residues generated on site. The simulation results have shown that for a typical MDF production line, the thermal energy demand for regular or standard grade of MDF is 989 kWh/[m.sup.3] using flue gas for direct fiber drying or 1168 kWh/[m.sup.3] when a hot air stream is used for fiber drying. The simulation results are in close agreement with the energy survey results of commercial plants with a discrepancy of--17 percent to +6 percent.
In commercial plants, medium density fiberboard (MDF) is manufactured through the following sequence of steps: log chipping, chip preheating and refining, fiber drying, mat forming and prepressing, mat hot pressing, and finishing. The operations in chip preheating and refining, fiber drying, and mat hot pressing consume large quantities of thermal energy and electricity. Electricity is commonly purchased, and thermal energy is supplied on site by combustion of the wood residues generated in the production process. Quantifying the thermal energy demand and distribution is important for managing energy supply and for potential reduction of the overall energy consumption. Even though most thermal energy at MDF plants is supplied from renewable sources, the drive to reduce the woodwaste contributing to these sources and the general drive to improve the utilization of energy makes energy management an important issue in the industry. The objective of the work presented in this paper is to establish an energy demand model to predict the thermal energy consumption and to analyze the thermal energy distribution in a typical New Zealand MDF production line to support efforts to improve energy efficiency.
In New Zealand, there are four MDF plants with a total of seven production lines. The MDF production capacity of each line ranges from 85,000 to 160,000 [m.sup.3]/yr with an average production of 125,000 [m.sup.3]/yr (MDF part II 1997). All seven production lines use one-stage fiber drying and pendistor vacuum mat forming. In the pressing step, two lines use batch (multiopenings) presses and the remaining five use continuous presses. All New Zealand MDF is manufactured from young radiata pine fibers, and these fibers are bonded with urea-formaldehyde (UF) resin, though melamine-formaldehyde (MF) or urea-melamine-formaldehyde (UMF) resins are occasionally used for specially required products. MDF is usually graded according to the panel density over a range from 450 to 880 kg/[m.sup.3] (Sturgeon 1992). Standard or regular grade of MDF panels has an average density of 720 to 730 kg/[m.sup.3] (French 2002, NPI 2005, Patinna 2005).
Because the thermal energy is supplied using the self-generated wood residues, its value is not typically quantified and thus no comprehensive model for the thermal energy demand and distribution has been found in literature. Existing models for the MDF processing have primarily considered how the panel properties are affected by pressing conditions in the hot-press (Carvalho et al. 2003, Thoemen and Humphrey 2003) and how the dry fiber production is affected by the drying operation parameters (Pang 2001). The thermal energy demand model presented in this paper is the first effort of its type for MDF production.
Modeling assumptions and considerations
Assumptions and considerations in the model development are as follows:
--The process uses one-stage fiber drying and continuous hot pressing.
--Moisture content (MC) and solid material load are expressed in oven dry base (od) and oven dry kg (odkg), respectively, unless specified.
--Pressure is expressed as absolute pressure.
--Radiata pine log is used as raw material, which is the dominant species for New Zealand plantation forests.
--UF resin is used as the binding agent.
--The base reference case point for the model is 120,000 [m.sup.3]/yr with 22.5 hours/day and 350 day/yr operation to allow for downtime and normal plant maintenance.
--Four MDF grades are defined for input of production data into the model according to density as follows: Thin grade--800 kg/[m.sup.3], Regular grade--725 kg/[m.sup.3], Light grade--600 kg/[m.sup.3] and Ultra-light grade--500 kg/[m.sup.3].
--Specific heat of wet chips or fibers is a combination of the specific heat of the oven dry wood and the moisture, thus it is a function of MC. It can be determined by using the formula of Pang et al. (1995).
--Mass balances are based on 5 percent bark on log, 2 percent and 4 percent chip fines respectively in chip screening and washing, 8 to 17 percent (w/w) solid UF resin being applied to the dry fibers with lower resin loading for the higher density grade, 4 percent panel trim off, 3 percent rejected panel, and 10 percent sander dust.
In the model development, the thermal properties of water, steam and wood are needed to quantify the heat requirement. The specific heat of water ([C.sub.pw]) of 4200 J/kg x K is used, which is an average value in the normal operating temperature range. Similarly, the steam average specific heat ([C.sub.pv]) of 2200 J/kg x K and the average latent heat of water vaporization ([DELTA][H.sub.wv]) of 2.1 MJ/kg are used for the temperature range of 100 to 180[degrees]C (Perry et al. 1998).
The thermal energy demand model for the MDF manufacturing process is developed based on energy and mass balances for materials of chip, fiber, resin, steam and moisture. Detailed processing steps in the three unit operations are considered:
1) Chip preheating and refining, which includes chip washing, screw feeding, preheating, and refining;
2) Fiber drying, which includes both the blowline and the fiber drying;
3) Mat hot pressing, which includes the mat forming, prepress and hot-press.
[FIGURE 1 OMITTED]
Preheating and refining
Chip washing and feeding.--Figure 1 illustrates the process of chip washing and feeding. General nomenclature for the figures and equations is listed in Table 1. It is essential that the chips are washed to remove stones, tramped metals, and sands or dirt before further processing (Allen et al. 1988). In the chip washing, the chips are thoroughly agitated in warm water (40[degrees]C) and pass through an upflow and settling section to remove the impurities. The waste water from the chip washing is cleaned in a hydro-cyclone to remove the solids before it is mixed with effluent from the screw feeder in the following step and recirculated back to the wash. The model assumes that the wood chip input is pure wood without any contamination, and thus the washed-out residue ([W.sub.wl], kg/h) can be regarded as a fuel with the same calorific value as the other wood residues in the process.
Heat consumed in the chip washing ([Q.sub.wash], J/h) is calculated from the known values of the chip load ([L.sub.ch], odkg/h) for washing, the specific heat of chips ([C.sub.pchip], J/odkg x K) at a MC M, the chip temperature before and after the chip washing (T and [T.sub.0], [degrees]C), and the normalized heat loss ratio ([[epsilon].sub.0]) in the washer.
[Q.sub.wash] = [L.sub.ch][C.sub.pchip]([T.sub.0] - T)/1 - [[epsilon].sub.0] 
After washing, the chips are heated to a temperature of about 35[degrees]C and transported by a chip pump to a dewatering device such as screw drainer above the chip hopper (Allen et al. 1988). In the washing, the chips are assumed to be fully saturated with the saturation MC ([M.sub.0]) varying with wood density. In the chip hopper, saturated steam of about 0.4 MPa is injected to heat the chips, which are then conveyed in a screw feeder to form a plug, in such a way that a large amount of moisture is squeezed from the chips. Plant observation shows that the chip outlet MC from the screw feeder ([M.sub.1]) is fairly uniform and constant at around 85 percent. The electrical energy ([N.sub.scf], W) required by the screw feeder is partly to convey the chips to the preheater and partly to compress the chips to form a plug. In doing so, frictional heat is generated that also heats up the chips. The chip temperature at the screw feeder outlet is typically regulated at 95[degrees]C. The quantity of steam consumed ([m.sub.1], kg/h) in the hopper is obtained from the energy balance in the chip hopper and the screw feeder to achieve the required temperature at the outlet:
[L.sub.0][C.sub.p0]([T.sub.1] - [T.sub.0])/1 - [[epsilon].sub.0] = [m.sub.1][[C.sub.pv] ([T.sub.steam1] - [T.sub.steam2]) + [DELTA][H.sub.wv]] + 3600[[alpha].sub.1][N.sub.scf] 
where [L.sub.0] (odkg/h) (= [L.sub.ch] - [W.sub.wl]) is the chip load to the screw feeder; [C.sub.p0] (J/odkg x K) is the specific heat of chips at the MC [M.sub.0]; [T.sub.0] and [T.sub.1] ([degrees]C) are the chip temperatures at the hopper entrance and the screw feeder outlet, respectively; [[epsilon].sub.1] is the heat loss ratio in the feeding system; [T.sub.steam1] and [T.sub.steam2] ([degrees]C) are the steam temperatures at 0.4 MPa and 0.1 MPa, respectively; and [[alpha].sub.1] is the coefficient of electricity-to-heat conversion in the screw feeder.
The hot water squeezed-out by the screw feeder includes the condensate of the steam injected and the water from the reduction of chip MC. It is estimated that the heat available in the effluent from the screw feeder is higher than the heat demand in the chip washing ([Q.sub.wash]). Therefore, recycling the screw feed effluent to the washing bath is sufficient to meet the heat demand in the chip washing. Consequently, the thermal energy demand in the washing and feeding is only from the injected steam to the hopper ([m.sub.1]) which can be determined from Equation .
Pre-heating and refining.--The process of preheating and refining is shown in Figure 2, in which the chips in plugs are fed from the screw feeder to a preheater or digester vessel. In the preheater, the chips are heated to a temperature of 170 to 180[degrees]C by saturated steam of about 1 MPa. At this temperature, the lignin and hemicelluloses of the chips are softened and the fibers can be easily separated in the downstream refining (Allen et al. 1988).
In the energy model, it is assumed that the preheater and the refiner have the same operating pressure and temperature. The effect of the small amount of wax added as a water repellent on thermal energy consumption is ignored.
In the preheater, chips with a specific heat [C.sub.p1] (J/odkg x K) at MC [M.sub.1] are heated from temperature [T.sub.1] to temperature [T.sub.2] ([degrees]C), which is the saturation temperature at the pressure of the injected steam. The heat required in the preheater is equal to that released from the condensation of the injected steam. Considering the chip temperature [T.sub.2] should be kept constant in the preheater and the refiner, the steam requirement ([m.sub.2], kg/h) can be determined using a heat loss ratio of [[epsilon].sub.2] in the system by:
[m.sub.2] = [L.sub.0][C.sub.p1]([T.sub.2] - [T.sub.1])/[DELTA][H.sub.wv](1 - [[epsilon].sub.2]) 
This is the minimum steam requirement in the preheater, however, in practice, an extra amount of steam ([m.sub.3a], kg/h) about 30 percent of [m.sub.2], is also supplied which flows to the refiner together with the heated chips.
[FIGURE 2 OMITTED]
In the refiner, electricity ([N.sub.rf], W) is provided to break the chips into fibers through mechanical action. In doing so, the electricity is converted to thermal energy and the refiner effectively acts as a steam generator (Allen et al. 1988). The steam generated ([m.sub.g], kg/h) is calculated by Equation  with [[alpha].sub.2] designating the electricity-to-heat conversion coefficient.
[m.sub.g] = 36000[[alpha].sub.2][N.sub.rf] - [DELTA][H.sub.wv] 
Some of the steam is recycled ([m.sub.g] [phi]) from the refiner to the preheater at a recycle ratio ([phi]) to reduce the fresh steam consumption ([m.sub.fresh], kg/h) to meet the total steam requirement ([m.sub.2], kg/h) in the preheater.
[m.sub.fresh] = [m.sub.2] - [m.sub.g][phi] 
A small additional amount of steam ([m.sub.3b], ~400 kg/h) is usually injected to the refiner in order to maintain the required pressure. By taking the preheater and the refiner as a single system, the total steam input (m, kg/h) to this system is given by Equation  and the steam going forward to the blowline ([m.sub.4], kg/h) is calculated from Equation .
m = [m.sub.fresh] + [m.sub.3a] + [m.sub.3b] 
[m.sub.4] = (1 - [phi])[m.sub.g] + [m.sub.3a] + [m.sub.3b] 
In the preheater, the chip MC increases with steam condensation and in the refiner, the fiber MC decreases with steam generation. Therefore, the outlet MC ([M.sub.2]) of the refined fibers is calculated by:
[L.sub.0][M.sub.2] = [L.sub.0][M.sub.1] + [m.sub.2] - [m.sub.g] 
In the above calculation it is assumed that the fiber production equals the chip load [L.sub.0] (odkg/h) in the preheater.
Blowline.--After refining, fibers are discharged through a valve into a blowline as shown in Figure 3. In the blowline, a velocity of up to 100 m/s for the mixture of fibers and steam can be achieved at the normal operating pressure of 0.4 to 0.5 MPa ([P.sub.in]). The UF resin solution ([L.sub.r] + [L.sub.water], kg/h) is added in the blowline where the turbulent flow provides desired mixing. No net energy is supplied in the blowline, but it is necessary to calculate the fiber MC change and the steam evaporation rate in the blowline for the input values in the subsequent fiber drying calculations.
Because of the highly turbulent steam-fiber flow, the pressure drop in the blowline is significant and can be estimated by (Perry et al. 1998):
[DELTA]p = f l/d [rho]/2 [u.sup.2] = 0.025 l/d u/2 [m.sub.4]/3.6/[pi][d.sup.2]/4 = 0.014 l x u x [m.sub.4]/[pi][d.sup.3] 
[FIGURE 3 OMITTED]
In Equation , f is the friction factor, which is estimated to be 0.025 for a stainless steel blowline pipe with a smooth surface; d is the inner diameter of the blowline pipe, which is about 100 ram; l is the length of the blowline pipe which is typically in the order of 50 m and u (m/s) is the steam-fiber velocity in the blowline.
At the entrance of the blowline, the steam becomes superheated once the pressure is reduced, and the superheated steam provides sensible heat to heat up the injected resin solution and to evaporate the water from the resin solution and the fibers. Heat loss ([Q.sub.loss], J/h) to the ambient atmosphere surrounding the pipe is also considered in the model. The steam at the blowline outlet reaches the saturation point corresponding to the outlet pressure ([P.sub.out], bar) and its temperature [T.sub.3] ([degrees]C) is readily taken from an empirical correlation fitted from a steam table in Perry et al. (1998):
[T.sub.3] = 50 [log.sub.10]([P.sub.out]) + 3.48[P.sub.out] + 96.5 
By considering the energy exchange between steam, fiber, resin, and moisture water, the energy balance can be established for the blowline as follows:
([L.sub.r][C.sub.pr] + [L.sub.water][C.sub.pw])[T.sub.ro] + ([L.sub.0][C.sub.p2] + [m.sub.4][C.sub.pv])[T.sub.2] = [[L.sub.r][C.sub.pr] + [L.sub.0][C.sub.p3] + ([m.sub.4] + [m.sub.evap])[C.sub.pv]][T.sub.3] + [m.sub.evap][DELTA][H.sub.wv] + [Q.sub.loss] 
in which [T.sub.r0] ([degrees]C) is the resin temperature at injection, [C.sub.p2] and [C.sub.p3] (J/odkg x K) are the specific heat of fibers at the inlet and the outlet of the blowline, respectively. The outlet fiber MC [M.sub.3] is related to the steam evaporation rate ([m.sub.evap], kg/h) for a given amount of resin solution based on the mass balance:
[M.sub.3] = [L.sub.water] + [L.sub.0][M.sub.2] - [m.sub.evap]/[L.sub.0] + [L.sub.r] 
Equations  and  can be solved to determine the outlet MC ([M.sub.3]) and the steam evaporation rate ([m.sub.evap]) as the other parameters are known from either the operation conditions or the refining.
Fiber drying.--Figure 4 shows the process of fiber drying. At the entrance to the dryer, the mixture of wet, resinated fibers and steam from the blowline is hot and depressurized to atmospheric pressure very quickly. In a short distance of 1 to 2 m from the dryer entrance, fast evaporation of moisture occurs and because of this, the fiber dryer is also referred to as a flash dryer. After the flash drying, a drying medium, either hot flue gas or hot air, is injected to mix with the fiber and steam. Along the length of the dryer, the wet, resinated fibers are dried to the required MC, normally 10 to 11 percent. The hot drying medium provides sensible heat for the drying and, at the same time, carries the evaporated vapor away from the fibers. The mixture of fiber, vapor and drying medium then goes into a cyclone for separation where the exhaust gas is vented from the top and the dried fibers are discharged from the bottom (Pang 2001).
[FIGURE 4 OMITTED]
The temperature of the drying medium is normally controlled between 120 and 180[degrees]C at the dryer entrance to prevent fiber claps and kinks that could result from more severe drying conditions (240 to 300[degrees]C) previously used in the older plants (Allen et al. 1988). The flow rate of the drying medium is controlled based on the final fiber MC, the production capacity and the target exhaust temperature (60 to 65[degrees]C).
The energy model for the drying process has considered two drying media, either the directly injected flue gas or the indirectly heated hot air. The flue gas drying medium is obtained from the exhaust flue gas from a thermal oil heater, normally at 380[degrees]C, which is then cleaned and mixed with ambient air to achieve the required drying temperature. For the hot air-drying, the ambient air is heated up to the required drying temperature by the 380[degrees]C flue gas in a typical heat exchanger.
Flue gas as the drying medium.--The heat balance Equation  is established for the fiber dryer based on the fact that the heat required is mainly used to reduce the MC of the fibers from [M.sub.3] to [M.sub.5]. The heat required is provided by the hot flue gas, the hot wet fiber and resin, and the steam with a dryer energy efficiency of [[eta].sub.dryer].
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
In the above equation, [C.sub.pf] (J/kg.K) is the specific heat of the dry fiber, [T.sub.4] ([degrees]C) is the drying temperature at the inlet and [T.sub.5] ([degrees]C) is the exhaust gas temperature. Due to the close contact between the fibers and the drying medium, the dry fiber has the same temperature as the exhaust gas at the outlet of the cyclone.
With the known parameters, the flue gas flow rate ([m.sub.fl], kg/ h) can be determined from Equation . Heat required ([Q.sub.fl], J/h) from the flue gas can then be calculated based on the specific heat of flue gas ([C.sub.pfl], J/kg K) at the average temperature of [T.sub.4] and [T.sub.5].
[Q.sub.fl] = [m.sub.fl][C.sub.pfl]([T.sub.4] - [T.sub.5]) 
The total heat demand ([Q.sup.fl.sub.dry], J/h) for drying can be calculated by considering a heat loss ratio in the flue gas cleaning and mixing step ([[epsilon].sub.clean]).
[Q.sup.fl.sub.dry] = [Q.sub.fl]/1 - [[epsilon].sub.clean] 
Hot air as drying medium.--The effective heat demand in the hot air-drying is equal to the heat demand in the flue gas drying ([Q.sub.fl], J/h) at the same production capacity. Therefore, airflow rate [m.sub.a] (kg/h) to the dryer can be calculated by:
[m.sub.a] = [Q.sub.fl]/[C.sub.pa]([T.sub.4] - [T.sub.5]) 
However, there is a heat efficiency factor ([[eta].sub.air]) for the heat exchanger where the air is heated to the required temperature by the flue gas. By considering this, the actual heat demand in the hot air-drying ([Q.sup.a.sub.dry], J/h) can be determined by the following equation in terms of the flue gas:
[Q.sup.a.sub.dry] = [m.sub.a][C.sub.pa]([T.sub.4] - [T.sub.0])/[[eta].sub.air] = [Q.sub.fl] ([T.sub.4] - [T.sub.0])/[[eta].sub.air]([T.sub.4] - [T.sub.5]) 
The higher heat demand in the hot air-drying compared to the direct flue gas drying can be explained by the fact that the ambient temperature ([T.sub.0]) is always lower than the exhaust temperature from the dryer ([T.sub.5]) and the heat efficiency in the heat exchanger is always less than 1.
Mat hot pressing
The dried fibers from the fiber dryer are usually stockpiled in a storage bin. From there the dried fibers are metered and conveyed to the forming station for mat formation followed by prepressing and final hot pressing. There is no heat demand in the mat formation and prepress. The mechanical work input has a negligible effect on the heat balance so it is not considered in the model and thus the mat is assumed to enter the final hot-press at ambient temperature. In order to shorten the hot pressing cycle, especially for the thick board, preheating is sometimes applied to elevate the mat temperature before the hot pressing (Maloney 1993). However, as preheating is not normally applied in the MDF plants in New Zealand, it is not considered in the model.
In a continuous press, the prepressed mat is continuously conveyed through the press between two thin stainless steel belts. The friction between the conveying bands and the hot platens is reduced by a carpet of thin rollers as illustrated in Figure 5 (Thoemen and Humphrey 2003).
Platens are commonly heated using hot oil in the New Zealand MDF industry, while steam or hot water heating with radio frequency preheating is used in the other countries (Maloney 1993). Hot oil, heated by flue gas in a separate heat exchanger, is pumped through the hollow space in the platens continuously.
The heat input to the fiber mat comes from the heated press platens and the exothermic curing reaction of the resin binder. It is observed that at the end of the press, the panel surface temperature is very close to the press platen temperature [T.sub.p], and the panel core temperature is at about 100[degrees]C. The 100[degrees]C core temperature must be maintained for roughly 30 seconds to properly cure the UF resin (Maloney 1993). An average value ([T.sub.6]) of the panel surface and the core temperature is taken as the overall MDF panel temperature after pressing. In this way, the heat ([Q.sub.mat], J/h) required to heat the fiber mat during the hot-press can be determined from the heat balance:
[Q.sub.mat] = [[C.sub.pf][L.sub.0] + [C.sub.pr][L.sub.r] + [C.sub.pw][M.sub.5]([L.sub.0] + [L.sub.r])] ([T.sub.6] - [T.sub.0]) 
[FIGURE 5 OMITTED]
The heat released from the resin cure reaction ([Q.sub.react], J/h) is equal to the polycondensation enthalpy of 84,130 J/kg for the UF resin with 1.5 percent (w/w) hardener (N[H.sub.4]Cl), which is the typical hardener content used in MDF production (Carvalho et al. 2003).
During the hot pressing, further drying usually takes place due to elevated vapor pressure within the panel thus the final panel MC is reduced to about 8 percent. Total water evaporation rate ([W.sub.evap], kg/h) in the hot pressing should include the MC change and the water formed by the resin polycondensation.
[W.sub.evap] = ([L.sub.0] + [L.sub.r])([M.sub.5] - [M.sub.6]) + [L.sub.r][W.sub.r] 
In order to determine the water released from the resin polycondensation ([W.sub.r], kg/h), an Arrhenius type kinetic model in differential form developed by Carvalho et al. (2003) was integrated and used in the model.
Heat demand ([Q.sub.press], J/h) for the press should include the energy needed to heat the MDF mat to the final temperature, heat released by resin curing, heat for water evaporation and the heat lost during the hot pressing.
[Q.sub.press] = ([Q.sub.mat] - [Q.sub.react] + [W.sub.evap] x [DELTA][H.sub.wv])/(1 - [[epsilon].sub.press]) 
Where [[epsilon].sub.press] is the heat loss ratio in the hot pressing, which can be as high as 50 percent of the heat required due to the fact that the heat is supplied from the platen surface to the mat in an open environment. This heat transfer is through three sequential stages: (1) heat conduction from the hot platen surface to the steel rods; (2) heat conduction from the steel rods to the stainless steel belt; and (3) heat transfer from the steel belt to the mat surface. Each of the three stages causes heat loss to the surrounding air by natural convection and radiation.
Coefficient fitting and validation of the model
The index value of the thermal energy demand (defined as the energy demand per unit volume of MDF panel) in kWh/[m.sup.3] for the regular grade is used to compare the energy demand data from commercial plants with the model prediction. The coefficients of electricity-to-heat conversion, heat loss ratio and energy efficiency have been adjusted according to the energy survey result of a commercial plant, Plant 1, which uses flue gas fiber drying. The coefficients were introduced based on the engineering principles and practices. These coefficients were fitted in a practical range by comparison between the predicted and measured energy consumption. Due to the limited number of plants for the data collection, more accurate coefficient fitting technique such as the least square root fitting was not used. Table 2 lists the fitting values of these coefficients used in the final version of the model. Figure 6 shows that the model thermal energy demand fits well with the Plant 1 data giving a total difference of 6 percent.
After the coefficient fitting, the model is validated using data obtained in further plant surveys. When hot air is used for fiber drying, Figure 7 illustrates the validation of the model compared to the energy demand in Plant 2 (energy audit data) (French 2002) and a typical plant reported by the Center of Advance Engineering in New Zealand (CAE 1996). Thermal energy demand simulated by the model is 2 percent to 17 percent lower than that reported values from the two plants. The lower predicted values are reflected in every plant unit operation. The discrepancy between the model prediction and the reported data are not considered to be significant since there is a generally wide range of values from the different plant cases. The result of Plant 2 was based on energy audit data and is expected to be more reliable for comparison (French 2002). The typical plant figure published by CAE (1996) is supposed to be the statistical result, however, there is no explanation of the original data source or collection method.
[FIGURE 6 OMITTED]
[FIGURE 7 OMITTED]
If a comparison is made between the model and the three plants (Plant 1, Plant 2 and Plant CAE), the model result has a discrepancy of -17 percent to +6 percent compared to the practical thermal energy demand data. Although the agreement of model and Plant 1 is clearly the result of parameter fitting, the accuracy of the model simulation is expected to be well within the accuracy of [+ or -]20 percent specified in a Level 2 Standard Energy Audits in accordance with the Australian and New Zealand Standard (AS/NZS 3598:2000). Plant 1 uses flue gas directly for fiber drying, and Plant 2 is the typical plant using hot air. Therefore, the proposed model in this work can be applied to simulate the thermal energy demand for both of the fiber drying cases.
Results and discussion
Energy demand by MDF grades
The energy model calculates the MDF plant heat demand based on unit oven dry weight of materials processed. With the same production volume, higher density grades result in higher production weights, thus consuming more thermal energy. The energy demand per unit volume of product is approximately proportional to the grade density.
The index values of energy demand and raw material demand are given in Table 3 and show that the production of lighter panel consumes less energy and also less log material per unit volume MDF panel. However, the lighter panel requires higher resin loading to achieve the required internal bonding property. In practice, the MDF production is a mixture of grades depending on the market demand. In this case, the model could be run separately for each grade for a particular period of production time. Regular grade is usually the majority of the products and thus this paper focuses on this grade in the following analysis.
Energy demand by unit operation
In the reference case with an annual production of 120,000 [m.sup.3] regular grade of MDF, the model predicts a total annual thermal energy demand of 119 GWh in flue gas heat when the flue gas is used for fiber drying. If hot air is used for the fiber drying, the total annual thermal energy demand increases to 140 GWh, 21 GWh more than the flue gas drying. Figures 8 and 9 show the energy demand for the three unit operations in the two fiber drying cases. Fiber drying is the biggest energy consumer, accounting for 45 percent of the total thermal energy consumed when flue gas is used or 54 percent if hot air is used. The higher energy consumption in the hot air-drying is due to the fact that more heat is lost in the flue gas venting at high temperatures (~150[degrees]C) after heating the air (French 2002). In some older plants, the flue gas exhaust temperature could be as high as 250 to 300[degrees]C although multiple economizers are used (Allen et al. 1988).
Thermal energy consumption in the fiber drying can also be reduced by decreasing the steam flow rate into the dryer from the refiner or by improving the dryer energy efficiency. Lower steam flow rate can be achieved by using a higher steam recycle ratio from the refiner to the preheater. However, this may result in a lower humidity in the dryer and some drying defects such as fiber collapse and resin precuring can result from too severe drying (Allen et al. 1998). A dryer energy efficiency of 60 percent is used in the model for the one-stage dryer which reflects the heat loss from the dryer walls and the cyclone. The dryer heat efficiency can be improved by using a two-stage dryer to recover the condensed moisture and heat from the first stage exhaust (Pang 2001).
Thermal energy distribution
The distribution of thermal energy demand as calculated from the model is illustrated in Figure 10 for flue gas fiber drying. Flue gas from a wood residue-fired furnace is the primary energy source. One common thermal oil system is used for the three energy users including steam generation, dryer heating and hot-press. Steam is used for chip feeding, preheating, and refining; flue gas itself is for fiber drying, and thermal oil is used for the hot pressing.
In the total thermal energy demand of 119 GWh (flue gas heat), 65 GWh is used for oil heating to produce 53 GWh thermal oil while the remaining 54 GWh is used for fiber drying. Along the hot oil flow path, 38 GWh of the thermal oil heat (48 GWh flue gas) is used to provide 34.7 GWh of total steam made up of 0.4 MPa (22.2 GWh) and 1 MPa (12.5 GWh), respectively, supplied to the chip hopper, and the preheater and refiner. The remaining 15 GWh oil heat is used for the hot-press to bring the mat to the temperature for resin curing. This thermal oil heat is provided from 17 GWh flue gas heat.
[FIGURE 10 OMITTED]
In the chip hopper, the 22.2 GWh steam at 0.4 MPa consumed is equivalent to 30.7 GWh flue gas heat. However, energy demand in the chip washing is not reflected in the diagram since it uses recycled hot water from the screw feeder after the chip hopper.
It is important to note that in the preheater and refiner, only 6.9 GWh of the 1 MPa steam is needed to heat up the chips to the refining temperature taking credit for the additional heat from the consumed electrical energy in the refining. The remaining 5.6 GWh of 1 MPa steam, which is used for maintaining an even temperature and pressure in the preheater and refiner, flows through the blowline to the dryer and is used in the fiber drying.
Fiber drying with directly injected flue gas consumes 46 GWh flue gas at a temperature of 160[degrees]C and 5.6 GWh of the steam from the refiner. Considering the heat loss in the gas cleaning and mixing, the actual energy input for the fiber drying, without the 5.6 GWh of steam heat from the blowline, is equivalent to 54 GWh of flue gas heat.
Thermal energy self-sufficiency
For production of 120,000 [m.sup.3]/yr regular grade MDF, wood residues generated are totaled at 25,691 odt/yr including bark, chip fines, trim-off, sander dust, and rejected panels. Sander dust and bark are the major residues accounting for over 50 percent of the total. Care should be taken on the wood residues MC when they are burned to generate thermal energy since many burners have difficulty with rapidly changing fuel moisture level. Fresh bark and fines from chip screen usually have a MC similar to the fresh log at about 120 percent. Fines from chip washing normally have a saturated MC of around 180 to 210 percent depending on the wood density. Trim off from the panel cross cutting, sander dust and rejected panels are nominally dry with a MC of 6 to 12 percent. Mixing the various residues may be necessary to achieve a homogeneous and steady MC for combustion or gasification.
The total wood residues have an energy value of 135 GWh/ yr based on a calorific value of 19 MJ/odkg (Baines 1993). They can generate thermal energy of 108 GWh/yr in flue gas heat using 80 percent thermal conversion efficiency in a furnace burning a fuel with an average 100 percent MC. This amount of energy is able to provide an energy self-sufficiency of 91 percent for flue gas fiber drying, or 77 percent for hot air fiber drying according to the model.
In a MDF production line, chip preheating and refining, fiber drying, and panel hot pressing are the three unit operations that consume thermal energy. Based on the processing technology of one-stage fiber drying and continuous hot pressing, a spreadsheet-based model has been established to quantify the thermal energy demand in the MDF production. The model can be used to examine the effects of various production capacities, product grades, operation times and fiber drying methods on thermal energy demand and distribution.
In the plants considered, the biggest difference in the thermal energy demand was how the hot flue gas from a wood residue furnace is used for fiber drying. One method takes the flue gas downstream from the thermal oil heater (now at 380[degrees]C), then cleans and quenches it with ambient air down to 160[degrees]C before entering the dryer for the fiber drying. The other method takes the flue gas from the same point in the flow sheet and sends it to a conventional heat exchanger to heat up a separate air stream for the fiber drying.
In a reference case of MDF production of 120,000 [m.sup.3]/yr, energy demand increases with the panel density, 676 kWh/[m.sup.3] for Ultra-light grade, 989 kWh/[m.sup.3] for the Regular grade and 1100 kWh/[m.sup.3] for the Thin grade using the flue gas fiber drying. When the hot air is used for the fiber drying in the production of Regular grade MDF, the total thermal energy demand is 1168 kWh/[m.sup.3] which is 179 kWh/[m.sup.3] (15%) more than the flue gas fiber drying. For this Regular grade case, the model predicts that combustion of the residues is able to provide 77 percent of its own thermal energy demand using hot air for fiber drying or 91 percent using flue gas for fiber drying.
Validation of the model using energy survey results from MDF plants shows the model is able to predict the energy demand with an accuracy of -17 percent to +6 percent, which is within the data variation of [+ or -]20 percent required in the energy audit (level 2) Standard of Australia and New Zealand. However, it is noted that the model calculation is based on a single grade MDF while in practical production, the product is a mixture of different grades during a 1-year period.
Alien, D., C. Saefstroem, and P. Wiecke. 1988. Design aspects of modern MDF plants. Appita 41 (2):93-96.
Baines, J.T. 1993. New Zealand Energy Information Handbook. Prepared and published with financial assistance from Energy and Resources Div., Ministry of Commerce, Fisher and Paykel Ltd, and Southpower.
Centre for Advanced Engineering (CAE) 1996: Part 6: Forestry Processing. Energy Efficiency, Industry and Primary Production, Vol. 2. Univ. of Canterbury, Christchurch, New Zealand. pp. 217-268.
Carvalho, L., M. Costa, and C. Costa. 2003. A global model for the hot-pressing of MDF. Wood Sci. and Tech. 37:241-258.
French, C. 2002. Energy Analysis of a Medium Density Fibreboard Production Facility. thesis of a Post Graduate Diploma of Sci., Dept. of Physics, Univ. of Otago, Dunedin, New Zealand.
Maloney, T. 1993. Modern Particleboard and Dry-process Fiberboard Manufacturing. Miller Freeman, San Francisco.
Nelson Pine Industries Ltd.(NPI) 2005. www.nelsonpine.co.nz/MDFProd.htm.
Pang, S. 2001. Improving MDF fibre drying operation by application of a mathematical model. Drying Tech. 19(8): 1789-1805.
--, R.B. Keey, and T.A.G. Langrish. 1995. Modelling the temperature profiles within boards during the high-temperature drying of Pinus radiata timber: The influence of airflow reversals. Inter. J. Heat Mass Transfer 38(2): 189-205.
Dongwha Patinna NZ Ltd.(Patinna) 2005. www.patinna.com
Perry, R.H., D.W. Green, and J.O. Maloney. 1998. Perry's Chemical Engineers' Handbook (7th Ed.). McGraw-Hill Inter. Editions.
Sturgeon, M.G. 1992. The future of panel products: Medium-density fibreboard. In: Pacific Rim Bio-Based Composites Syrup., FRI Bulletin No. 177. pp. 42-50.
MDF Part II. 1992. World MDF capacity update part II: Asia-Pacific and the rest of the world. Wood Based Panels Int'l. 17(5): 12-13.
Thoemen, H. and P. Humphrey. 2003. Modelling the continuous pressing process for wood-based composites. Wood and Fibre Sci. 35(3): 456-469.
The authors are, respectively, Research Engineer and Associate Professor, Dept. of Chemical and Process Engineering, Univ. of Canterbury, Christchurch, New Zealand (firstname.lastname@example.org; email@example.com), and Director, Delta S Technologies, Port Chalmers, Dunedin, New Zealand (firstname.lastname@example.org). The study is funded by the New Zealand Foundation for Research, Sci. and Technology. The authors greatly appreciate Mr. Philip Wilson in Nelson Pine Industries Ltd., Mr. Alistair Allan in Dongwha Patinna NZ Ltd., and Mr. Dave Allen in Metso Panelboard for their valuable help in conducting this study. This paper was received for publication in November 2006. Article No. 10269.
Table 1.--Guide to variables used in equations and figures. Variable designation Explanation [L.sub.i] Dry mass flow rate of the wood chips or fiber at point i in the process in unit of odkg/h (oven dry kg per hour) [m.sub.i] Mass flow rate of the water present at point i in the process in unit of kg/h [M.sub.i] MC of the fiber material at point i in the process (percent dry basis) [Q.sub.i] Heat flow at point i in the process (J/h) [T.sub.i] Temperature of the process stream at point i in the process ([degrees]C) Table 2.--Coefficients, heat loss ratios and thermal energy efficiencies used in the model. Value based on energy Item Symbol supplied (%) Coefficient of electricity-to-heat by the [[alpha].sub.1] 40 screw feeder Coefficient of electricity-to-heat by the [[alpha].sub.2] 70 refiner Heat loss ratio in chip washer [[epsilon].sub.0] 20 Heat loss ratio in chip hopper [[epsilon].sub.1] 35 and screw feeder Heat loss ratio in preheater [[epsilon].sub.2] 20 and refiner Heat loss ratio in flue gas [[epsilon].sub.clean] 15 cleaner and mixer Heat loss ratio in hot pressing [[epsilon].sub.press] 50 Thermal efficiency of fiber [[eta].sub.dryer] 60 dryer Thermal efficiency of heat exchanger for flue gas heating up [[eta].sub.oil] 85 thermal oil for thermal oil raising [[eta].sub.steam] 85 steam for flue gas heating up air [[eta].sub.air] 85 for thermal oil heating up [[eta].sub.platen] 85 platens Table 3.--Index value of energy and raw material demand with grade (flue gas for fiber drying). Ultra- MDF grade Thin Regular Light light Thermal energy (kWh/[m.sup.3]) 1100 989 815 676 Log ([m.sup.3]/[m.sup.3]) 2.27 2.01 1.62 1.31 UF resin solid (kg/[m.sup.3]) 65 79 81 80 Figure 8.--Annual energy demand by unit operation for production of 120,000 [m.sup.3]/yr regular MDF using flue gas for fiber drying. Fiber drying 54GWh, 45% Hot pressing 17GWh, 15% Preheating & refining 48GWh, 40% Note: Table made from pie chart. Figure 9.--Annual energy demand by unit operation for production of 120,000 [m.sup.3]/yr regular MDF using hot air for fiber drying. Fiber drying 54GWh, 54% Hot pressing 17GWh, 12% Preheating & refining 48GWh, 34% Note: Table made from pie chart.