EVALUATION OF A BIOMASS DRYING PROCESS USING WASTE HEAT FROM PROCESS INDUSTRIES: A CASE STUDY Hanning Li, Qun Chen, Xiaohui Zhang, Karen N Finney, Vida N Sharifi, Jim Swithenbank Sheffield University Waste Incineration Centre (SUWIC), Department of Chemical and Biological Engineering, University of Sheffield, Sheffield, S1 3JD, UK ABSTRACT Large amounts of low-grade waste heat from a range of process industries is widely available, in both their cooling water, at about 90°C, and their flue gases, at 250-400°C. Depending on the temperature of the heat available, these could be used for a number of applications, such as district heating in cities, cooking in the food industry and for the drying of biomass fuels. Dry biomass provides considerable benefits for combustion and gasification compared to their initial, ‘raw’ state, such as increased boiler efficiency, lower flue gas emissions and improved boiler operations. Drying is however an energy-intensive pre-treatment. An evaluation of the overall process is therefore required before more detailed research, design and construction work can be carried out. This paper presents the results of an investigation in which the integration of a drying process into a power station fuel system has been explored. Waste heat from a process industry plant (with an output of 100 MW) was utilised as the heat source for biomass drying. The biomass – in this case, pine wood chips at 60wt% moisture – was dried and could then be provided as the input fuel for a subsequent 40 MW power station; this can be combusted in a boiler at an improved thermal efficiency. The drying process evaluated consisted of a belt conveyor as the dryer and either the waste flue gases or superheated steam (generated from the hot cooling water) as the heat source. The results show that either source of low-grade waste heat exiting the industrial process plant could be successfully used to dry the wood. In the selection of the heat source (flue gases or superheated steam), flue gas usage would result in lower capital costs (~€2.5m), but the environmental issues must also be considered, such as any pollutant emissions and contaminated waste water that are generated. Superheated steam can combine short drying times, good heat recovery and environmental protection, but the higher capital costs (~€3m) can be a significant issue. A 3-4 year return on the initial investment was calculated for these technologies, although profitability was sensitive to the price at which the fuel can be sold. Key Words: low-grade waste heat; biomass; belt dryer; process industry.

1. INTRODUCTION Over the past few decades, the combination of a number of issues has meant that developing sustainable and renewable energy sources and improving the efficiency of systems using thermal energy have become increasingly important. The depletion of natural resources, due to rapid fossil fuel consumption, and environmental issues, like climate change and acid rain, are just some of these problems. Biomass can be used as a form of renewable energy, for both heat and power generation through thermochemical treatments, such as combustion and gasification. The term ‘biomass’ refers to both energy crops (plants grown specifically to be used as a fuel) and wastes/by-products, such as forestry residues, sawdust and a range of other agricultural and commercial/industrial wastes, which can all be utilised for energy production in much the same way as coal. Biomass is usually combusted on a grate (fixed- or moving-bed) or in a fluidised-bed boiler. The moisture content of biomass is typically high, often varying between 50wt% and 63wt%, depending on the season, weather and type of biomass. Typical lower heating values (the net calorific value) of dry biomass fuels are about 18.5-21.0 MJ/kg (Kiranoudis, et al., 1995). Unfortunately, the energy needed for the evaporation of water in a combustion boiler cannot be utilised in the power generation process, since the temperature level of the latent heat is too low. An initially low level of fuel moisture however could recover much of the energy used during combustion for water evaporation. It would also be beneficial for decreasing the dimensions of the boiler and reducing the emissions of unburned solids. Biomass with lower moisture contents could also minimise or eliminate other combustion control problems caused by fluctuations in the fuel moisture. Nonetheless, drying biomass is an energy-intensive process and can easily account for up to 15% of industrial energy utilisation (Chua, et al., 2001). Consequently, in many 1

industrial drying processes, a large fraction of energy is wasted (Ogura, et al., 2005). Energy management is therefore an essential part of any drying process and energy conservation can significantly lower the overall operating costs (Ho, et al., 2001). This paper investigated a biomass drying process using low-grade waste heat as the heat source. The heat source (100 MW) consisted of either waste flue gases at 250-450°C or hot water at 90°C, both exiting an industrial process plant. After drying, the lower moisture content pine wood chip fuel was then supplied to a 40 MW power generator. Two alternative drying systems, flue gas drying and steam drying with a water pre-heating process, were compared to assess the differences in energy consumption. A continuous belt dryer with a heat exchanger (if steam drying was used) was considered as the dryer configuration in both systems. The dryer design mainly consisted of the determination of various sizing and operational variables. The evaluation of specific process variables for each design was carried out using economic criteria. Both the capital and running costs were included in the evaluation and the profitability was assessed by determining the net present value (NPV).

2. INDUSTRIAL DRYERS FOR BIOMASS DRYING The dominant combustion technique for biofuels in the 1970s and 1980s was grate firing. This type of boiler can handle fuels with varying levels of moisture, but ideally one of 30-40% should be used (Wimmerstedt, 1995). Since the 1970s, fluidised-bed boilers have generally replaced grate-firing as a combustion technique (Huhtinen and Hotta, 1999). Compared with grate firing, fluidised-bed boilers are a more suitable method of combustion for moist biofuels. The use of fuels with a high moisture content however decreases the overall energy efficiency of the power plant and reduces the boiler capacity to such an extent that it becomes reasonable to install a dryer in combination with the boiler. In the 1970s and 1980s, industrial dryers tended to be direct flue gas dryers (Holmberg, 2007). Flue gases were either taken directly from the boiler or generated in a separate flue gas burner. The most common dryer types are rotary dryers, flash dryers, fluidised-bed dryers and belt dryers. Typical performance data for these are presented in Table 1, along with other considerations in Table 2. Table 1: Typical range of design parameters and performance data for various dryers. Sources: Bruce and Sinclair (1996); Amos (1998); Barré and Bilodeau (1999); van Deventer (2004); Holmberg (2007) Dryer Type Evaporation Rate (t/h) Drying Temperature (°C) Capacity (t/h) Feed Moisture at Inlet (%) Moisture Discharge (%) Feed Moisture at Outlet (%) Pressure Drop (kPa) Optimal Particle Size (mm) Maximum Particle Size (mm) Thermal Requirement (GJ/t-evaporation)

Rotary 3-23 200-600 3-45 45-65 10-45 2.5-3.7 19-50 25-125 3.0-4.0

Flash 4.8-17 150-280 4.4-16 45-65 10-45 12 7.5 0.5-50 2.7-2.8

Belt 0.5-40 30-200 45-72 15-25 25 0.5 1.26-2.5

2.1 Dryers As outlined above, there are several different types of dryer that are available. Rotary dryers, for instance, are the most common type for biomass applications and have low maintenance costs. Their robust and simple construction combines flexibility with reliability, enabling this type of dryer to operate under the most arduous conditions. They can handle a vast range of materials and are less sensitive to particle size, as shown in Table 2. The material moisture however is hard to control in rotary dryers because of the long lag time (Fredrikson, 1984). Though their design does permit the use of the highest possible drying/operating temperature (Table 1) and they can accept hot flue gases, this poses a considerable fire risk (Table 2); this also means that they require a lot of space – the most of all dryer types. Most dryers have outlet temperatures higher than 100°C to prevent the condensation of acids and resins.

2

Table 2: Considerations when choosing a dryer. Sources: Bruce and Sinclair (1996); Amos (1998); Barré and Bilodeau (1999); van Deventer (2004); Holmberg (2007) Dryer Type Requires Small Particles Heat Recovery Fire Hazard Air Emission Steam Use

Rotary none difficult high medium yes

Flash yes difficult medium high none

Belt none easy low low yes

Fluidised-Bed none easy medium medium yes

Flash dryers are able to dry biomass rapidly, owing to the easy removal of free moisture. Wet material is mixed with a stream of heated air (or other gas), which conveys it through a drying duct where high heat and mass transfer rates rapidly dry the product. Flash dryers require smaller biomass particle sizes to suspend and transport the biomass by the fluid stream alone (Table 2). Gas temperatures tend to be slightly lower than for rotary dryers. Flash dryers are much more compact than rotary dryers, but have higher installation costs (Fredrikson, 1984). They also have high blower power costs in addition to the heat requirements for drying. Flash dryers have a lower fire risk than rotary dryers due to the shorter retention times and lower operating temperatures (Table 1 and 2). They can be used to dry most types of biomass. In belt dryers, the feedstock is spread on a moving perforated conveyor to dry the material in a continuous process. Fans blow the drying medium through the belt and biomass material. Belt dryers are very versatile and can handle a wide range of materials. They are now frequently used in low temperature operations to save energy, reduce air emissions and minimise fire hazards (Tables 1 and 2). Dryers can also be classified into fixed-bed and fluidised-bed designs according to the hot air velocity flowing through the bed. In a fluidised-bed dryer, the hot air flows through the bed at a velocity sufficient to support the weight of particles in a fluidised state. Bubbles form and break within the bed and as a result, there is a high volume of gas in contact with the biomass particles, leading to high heat and mass transfer rates, providing fast evaporation (Table 2). 2.2 Selection of the Dryer Belt dryers are better suited to take advantage of low-grade and waste heat because they operate at lower temperatures than rotary dryers (Table 1). Rotary dryers, for example, typically require inlet temperatures of 260°C, but more optimally operate around 400°C. In contrast, the inlet temperature of a belt dryer, such as a commercially-available vacuum dryer, can be as low as 10°C above the ambient temperature, although more typically they operate at higher temperatures, between 90°C and 200°C. Because of their lower temperature operation, fire hazards and emissions to the air are lower for belt dryers (Table 2). Using steam to dry moist fuels has recently attracted much interest for a number of reasons: the high energy efficiency, low fire hazard and better environmental control. Steam drying is mostly done using belt feeders or fluidised-beds (Table 2). The superheated steam in the dryer provides the thermal driving force necessary to evaporate the moisture in the fuel. Generally, the wet material is mixed with enough superheated steam to dry the material and end with saturated steam. There are disadvantages to this process however; these include the requirement of a small particle size to ensure good mixing with the steam, the high capital costs incurred for a stainless steel pressure vessel and wastewater treatment issues. In the design of a drying process, fire safety and emission issues should be considered. Fire safety refers to precautions that are taken to prevent the likelihood of a fire. Fires start when a flammable and/or combustible material with an adequate supply of oxygen are heated to their ignition temperature. Biomass generally has an auto-ignition temperature of 260-280°C. In most cases, air drying poses a potentially high fire risk, because of the high amount of oxygen in the air supply. Flue gas dryers can operate at higher temperatures than air dryers, because flue gases contain lower amounts of oxygen (Amos, 1998). Compared with air or flue gas dryers, superheated steam drying processes have an even lower fire risk because no oxygen is present. There are additional fire risks if the dried biomass is heated above its ignition temperature (Amos, 1998). As the high temperatures used in these dryers are a fire hazard, one effective method is to reduce the operating temperature; a temperature of less than 100°C could significantly minimise the likelihood of a fire. 3

The exhaust gas from a biomass dryer mostly contains sulphur dioxide, carbon dioxide, carbon monoxide, hydrocarbons and suspended particulate matter as pollutants. SO2, CO2 and CO can be removed by absorption processes before the exhaust is released to the atmosphere and particles can be removed in part by cyclones and wet scrubbers. All types of woody material contain volatile organics that may be emitted together with the water vapour. The emissions from biomass drying are greatly affected by the drying temperature, especially when it exceeds 100°C (Holmberg, 2007). Below 100°C, emissions are reported to be low (Spets and Ahtila, 2004). Exhaust gases or unclean condensates must be treated after the dryer if they contain high concentrations of the above emissions; this increases the overall drying costs. The drying temperature when using flue gases should be controlled so that it remains below 100°C to reduce the gas treatment costs, even though the temperature of flue gas is normally much higher than this. In steam drying, contaminated condensates include aerobic biological organisms, organic compounds, organic carbon and non-condensable components, such as CO2, H2, CO, CH4 and C2-C4 compounds, which require removal from the gas stream before release to the atmosphere (Bruce and Sinclair, 1996). In addition to the safety and environmental considerations, the selection of a dryer should also take into account the water evaporation rate, biomass properties (such as size), operating temperature and the availability of heat resources. Table 2 summarized some considerations in choosing the dryers. The significant advantages of rotary dryers are that they are less sensitive to material size, operate at high temperatures to reduce drying time, have a wide range of evaporation rates and are easy to install. The main drawback is the greater fire risk, due to this high operating temperature. Gaseous emission from this type of dryer also need to be highly controlled and heat recovery is difficult. Flash dryers on the other hand are more compact and easier to control, but require a small particle size, thus reducing the size of the material may be beneficial for drying, although this is an energy-intensive operation. Flash dryers can also be used in high capacity water removal applications. Belt dryers are used in low temperature operations with reduced fire risks, fewer gaseous emission and low energy consumption. The advantages of belt dryers over the other types for biomass drying hence means that the feasibility of their application was assessed herein. 2.3 Drying Rate In the operation of a belt drying process, air or steam that is used as the drying media flows through the solid bed and comes into contact with the surface of the material (fuel). This convective drying process removes water from the surface of the material whilst increasing the temperature of the fuel, because the temperature of the drying media stream is higher than that of the fuel. During the convective drying process, two distinct phases can be distinguished in the material (Gigler, et al., 2000). During the first stage – the constant drying rate period – the surface water on the material is removed. In the second phase – the falling drying rate period – internal diffusion of the water to the surface of the material takes place. These physical phenomena have been described by various models (Fyhr and Rasmuson, 1997; Gigler, et al., 2000; Tang, et al., 2004). In the first kind of model, mass and energy balances are formed to describe the convective mass and heat transfer between the solid surface (fuel) and the gas stream (drying media). In these models, partial differential equations have a good physical basis to predict the drying process of a material with appropriate drying models, given the drying stream conditions and the mass and characteristics of the material. In general, there are three equations needed to predict the drying properties of wood: those for the stream temperature, wood temperature and drying rate. In the second type of model, the drying rate is described by the characteristics of the material, such as porosity, hardness, pore size and particle size, among others. The drying rate is generally determined by experimental observations and then developed into an empirical relation. The drying rate curve can be also used for estimating the residence time of materials in the dryer. Various studies have investigated fuel drying, many of which have focussed on wood fuels. Sheikholeslami and Watkinson (1992), for example, explored the water evaporation rate from wood-residue fuel, mainly bark, comparing air and superheated steam drying. The maximum drying rate was obtained after a short drying time and then the drying rate rapidly decreased. Comparing the effect of temperature on both the superheated steam drying and air drying techniques, the maximum drying rate was much higher with the superheated steam than with the relatively dry air at temperatures above approximately 180°C, while the relationship was reversed below this point. The maximum drying rate represents the initial drying rate, which identifies the optimal operating conditions. The results indicated that in view of the drying rate, air drying is the preferred option to accelerate the drying process at temperatures below 180°C, whilst steam drying will significantly improve the drying rate at temperatures beyond 180°C.

4

Later, Fyhr and Rasmuson (1997) investigated the drying rates of pine and spruce woods at various particle sizes and operating temperatures in a superheated steam system. They found that drying pine wood takes less time than spruce wood, mainly due to the internal structure of the pine being more permeable than the spruce. They also assessed the effects of solid size (L, the longitudinal particle length, in m, and T, the thickness, also in m) on the total drying residence time (TDT), which can be estimated by: TDT1 = TDT 0

T 0 L1 T1 L 0

(1)

Gigler, et al. (2000) carried out drying experiments for willows chips and simulated drying for the same chips using air flow. At the beginning of the drying process, convective heat transfer dominates, leading to a rapid drop in the fuel moisture content. Following this, the diffusion of water inside the solid becomes significant, slowing down the drying rate. An increased chip size lengthens the drying time. Most recently, Holmberg (2007) studied the drying rates of pine wood bark in a bed of variable height using air as the drying medium. It was found that an increased drying temperature significantly reduced the drying time and hence accelerated the drying process. Furthermore, it was concluded that the bed must be deep enough so that the drying air at the exit has reached its saturation point. Increasing the bed height still decreased the dryer size but the influence on dimensions was not as significant as for thin beds. Bed heights between 0.2m and 0.8m are generally selected for conveyor dryers.

3. DESCRIPTION OF DRYING SYSTEM: A CASE STUDY The UK faces the combined challenges of maintaining secure energy supplies and reducing carbon dioxide emissions to address climate change. Biomass can provide an alternative source of energy, replacing coal, to alleviate, at least in part the above issues. Accordingly, a 40 MW power station has been proposed for Sheffield, UK, using waste wood as the fuel source. Fresh wood chips can have a considerable moisture content, often up to 70%, which would significantly reduce the thermal efficiency of the combustor if they were used directly. Reducing the moisture content through drying was thus required to improve efficiency. As stated above, biomass drying is an energy-intensive process, so utilising waste energy can make this more efficient and reduce the overall energy demand. In a survey of process industries near Sheffield, it was found that the waste heat contained in flue gases, with a temperature of 250-450°C, could be recovered and used as an energy source with the greatest potential for biomass drying. 100 MW of energy can be provided in the form of waste heat: 60% available thermally as hot water at 90°C, and the rest as hot flue gas. Figure 1 shows a schematic diagram of biomass drying using waste heat from a process industry plant. The mass flowrate of the 90°C hot water exiting the plant has been estimated to be 737 t/h. The mass and volumetric flowrates at the flue gas exit are shown in Table 3. 40% flue gas 250-450°C 60% hot water 90°C

heat source

100 MW combustor

fresh biomass

dryer

dried biomass

40 MW combustor

Figure 1: Schematic of the biomass drying process integrated into a power station. 5

Table 3: Flowrates of the flue gas exiting the process industry plant at various temperatures.

Flowrate Mass (kg/s) Volumetric (Nm3/s) Volumetric (×105 m3/h)

250 179.71 139.29 9.61

Temperature (°C) 300 350 400 146.24 123.12 106.19 113.35 95.43 82.31 8.56 7.84 7.30

450 93.26 72.29 6.89

To estimate the usage of the waste heat required for biomass drying, two process configurations were proposed and compared. The first was the direct use of the flue gas as the heat source, without the need for a heat exchanger. The temperature of the inlet gas flow in this first drying option was 250-450°C. Figure 2 shows a flow diagram for this adiabatic drying process. The second configuration utilises the flue gas to raise the hot water to superheated steam at a desired temperature via a heat exchanger. The generated steam can then be used as the energy source for drying. Figure 3 outlines this drying process and the associated pre-heating. The inlet superheated steam in the second configuration was around 150-180°C at 1-2 bars. The drying temperature is usually 20°C lower than the temperature of the heat source. These waste thermal energy sources were then evaluated for supplying heat to dry the biomass. The capital and operation costs, as well as the profitability of using a conveyor-belt dryer were subsequently estimated to provide information for the industrial design and construction of this drying process. flue gas feed

wetted flue gas DRYER

wood feedstock

dried wood

Figure 2: Flow diagram of the adiabatic drying process using waste flue gases as the heat source. recycled steam steam 90°C water

PRE-HEATER

saturated steam DRYER dried wood

wood feedstock flue gas feed

Figure 3: Flow diagram of the adiabatic drying process using superheated pressurised steam generated from the hot cooling water as the heat source, heated by the waste flue gases in a pre-heater.

Direct use of the flue gases for drying can mitigate the need for a heat exchanger, as shown. A problem of direct drying though is commonly believed to be contamination of the biomass, but this should not be an issue, since the biomass fuel is to be combusted. The use of the flue gas at a high temperature (250-450°C) could increase the drying rate, meaning that a dryer with smaller dimensions could be used. These higher drying temperatures however can cause problems, such as increasing both the amount of gaseous emissions and the fire risk, as identified above (Amos, 1998; Spets and Ahtila, 2004; Holmberg, 2007). Thus the design of the dryer must take into consideration that the temperature of flue gas at the outlet of the dryer should be less than 100°C (Amos, 1998). The use of steam as the drying medium, however, represents a more valuable energy source, because by allowing the steam to expand in a turbine, it would be possible to also recover mechanical work (electricity) from the steam (Bruce and Sinclair, 1996). Furthermore, the construction of the associated heat exchanger for this dryer arrangement would increase the capital costs. The flue gas or steam demand for drying depends on the type of dryer and many drying parameters. The waste flue gas from the industrial plant could provide sufficient heat for two alternative drying approaches. Industrial conveyor-belt dryers are the most popular type for removing moisture from agricultural products. As discussed in the review of industrial dryers above (Section 2), belt dryers are able to use low-temperature heat sources to achieve biomass drying. Moreover, the utilisation of a belt dryer will reduce the fire risk and minimise emissions, compared to other dryers if the exit temperature of the flue gas is designed to be below 100°C in the case of direct drying. The interior of this type of dryer is illustrated in Figure 4. 6

flue gas out IN wood at a high moisture content and specific temperature

OUT wood at a lower moisture content and specific temperature flue gas in

Figure 4: Side view of a continuous cross-flow dryer.

4. ESTIMATION OF FLUE GAS USAGE FOR BIOMASS DRYING The mass flowrates of flue gas in the two process configurations were determined according to the heat and mass balances for the heating medium and the biomass streams in a steady-state adiabatic process. Firstly the dryer capacity was estimated and then the flue gas usage and superheated steam requirements were calculated for the different drying configurations.

4.1 Capacity of the Dryer The solid biomass – white pine wood chips – had an initial moisture content of around 50-60wt%, as shown in Table 4. The dried fuel was to be the energy input for a 40 MW power plant. The heating value of this fuel was 16.66 MJ/kg on a dry basis. The mass flowrate of the dry biomass was calculated from the power input requirement and biomass heating value. The evaporation rates of water from the biomass were then evaluated (Table 4). Table 4: Evaporation rates of water from the solid biomass. Fuel Moisture Content (wt%) Initial, Final 60, 10 60, 20 60, 30 50, 10 50, 20 50, 30

Evaporation Rate kg/s t/h 3.3339 3.0005 2.5718 2.1337 1.8003 1.3716

12.0019 10.8017 9.2586 7.6812 6.4810 4.9379

4.2 Estimation of the Flue Gas Requirement The amount of flue gas required was estimated based on the heating rate of the flue gas that can evaporate water from the biomass. It is common practice to assume that within the interior of a dryer, the drying stream follows an adiabatic process. The enthalpy of the flue gas at the entrance of the dryer (Hf,in, in kJ/kgdry flue gas) was equal to that at the outlet of the dryer (Hf,out): H f, in = H f, out

(2)

The enthalpy of the flue gas with a specific water content could thus be estimated using the humidity data for air containing water: H f = (C p ,air + C p , water Humf ) × Tf + H latent × Humf

(3)

where Cp,air is the specific heat of the air, Cp,water is the specific heat of the water vapour, Humf is the humidity of the flue gas, Tf is the flue gas temperature and Hlatent is the latent heat of water. The humidity of the flue gas can be calculated in terms of the saturation pressure: Hum f =

M r ,water ϕ Psat × M r ,air P − ϕ Psat

(4)

7

where Mr is the molecular weight or mass (in this case of the water and the air), φ is the relative moisture, P is the pressure and Psat is the saturated pressure (both in mmHg). This saturation pressure can be estimated with the Antoine equation:

log10 (Psat ) = A −

B C+T

(5)

where the parameters A, B, and C are constants, selected according to temperature, either above or below 100°C. During the drying process, a known amount of water was removed from the solid biomass over a certain period of time – Wevap, the water evaporation rate (kg-water/s). A corresponding amount of flue gas must therefore have removed the water vapour from the dryer at the specified humidity. The flowrate of the flue gas balanced the mass of water evaporated from the material in the dryer. The mass flowrate of the flue gas (Gf in kg/s) was then determined according to the water removal rate at a given humidity change (the difference in humidity between the inlet and outlet, Humf,in - Humf,out) in the flue gas: Gf =

Wevap Hum f ,in − Hum f ,out

(6)

5 5 3 3 FlueFlue Gasgas Volumetric Flowrate /hr) volume flow rate (10 (x10mM /h)

The mass flowrate of the flue gas determined from Equation 6 was then used to evaluate the availability of the supplied flue gas. Figure 5 shows the volumetric flowrate of the flue gas at varying flue gas temperatures for different final fuel moisture contents. The maximum flowrate to obtain the required amount of flue gas was 2.3×105 m3/h, which was sufficient to dry the biomass. Figure 5 also demonstrates that a higher flue gas temperature could reduce the loading of flue gas in the dryer. 2.6 2.4 2.2 2

Final MC final moisture content (wt%-wet)

1.8 1.6

1010 wt%

1.4

2020 wt% 3030 wt%

1.2 1 150

250

350

450

550

Flue Gas Temperature (°C) o

Flue gas temperature ( C) Figure 5: Flue gas flowrates at varying flue gas temperatures for drying pine wood chips, where the initial moisture content of the fuel was 60wt%.

4.3 Estimation of the Superheated Steam Requirement Superheated steam was the alternative option for drying the biomass. In the absence of other gas species in the flue gas, water vapour is a major component in the heating medium. The mass flowrate of the steam in the dryer was estimated based on an adiabatic process, as for the flue gas case above, limited by saturated steam at a given temperature. In the steam drying process, the heating source was steam that can be partially generated by using the hot water at 90°C, but an integrated pre-heating process using the high temperature flue gas can be used as an additional source, as shown in Figure 3. Thus, the flowrate of the flue gas also needed to be evaluated in the thermal balance to determine the energy required to raise the 90°C hot water to the desired superheated steam temperature of 140-180°C. Figure 6 shows the variation in the mass flowrate of steam with steam temperature at different final moisture contents. As expected, an increased steam temperature results in a reduction in the flowrate of the steam required. For the same final fuel moisture content, a fuel with an initially high moisture content required a 8

higher flowrate of steam, and thus, for the same initial moisture, a fuel with a higher final moisture content required a lower steam flowrate. The steam recycle ratio (R) had no effect on the required flow rates since the recycled steam was mixed with the generated steam before entering the dryer. However, varying the recycle ratio significantly affected the operation of the pre-heater, because an increased recycle ratio reduced the generation of steam, consequently lowering the requirement of flue-gas usage, as shown in Figure 7. It is interesting to note that at a high recycle ratio, i.e. R=0.75 (75%), flue gas usage is negligibly affected by the flue gas temperature for steam generation in the pre-heater and steam temperature in the dryer. Figures 6 and 7 also demonstrate that the maximum flowrate of flue gas was about 2.4 × 105 m3/hr and the maximum flowrate of the 90°C hot water was 12 t/h. The maximum amount of waste low-grade energy in terms of the available flue gases (up to 9.6 x 105 m3/hr) and 90°C hot water (737 t/hr) that can be supplied would be sufficient to generate enough steam at a temperature of 140-180°C.

SteamFlowrate flow rate(t/h) Steam (t/hr)

14.00 12.00 10.00 Final content MC final moisture (wt%-wet)

8.00

10 wt% 10 20 wt% 20 30 wt% 30

6.00 4.00 2.00 120

140

160

180

200

o

Steam SteamTemperature temperature(°C) ( C)

G(105xM3/h) Flue Gas Volumetric Flowrate (105 m3/hr)

Figure 6: Steam flowrates at different steam temperatures for drying wood, at an initial moisture content of 60wt% and a flue gas temperature of 250°C. 3.00 2.50 2.00

steam recycle ratio

1.50

R=0% R=0

1.00 R=0.5 R=50%

0.50

R=0.75 R=75%

0.00 200

250

300

350

400

450

500

Flue Gas Temperature (°C) o

Figure 7: Flue gas flowrates required for generating steam at various flue gas temperatures and for different steam recycle ratios, where the steam temperature was 140°C, the initial fuel moisture content was 60wt% and the final moisture content was 10wt%.

5. THE COST OF DRYING The overall costs of biomass drying consist of both the capital costs (Costcapital) and the running costs (Costrun). The capital cost is generally considered to primarily be composed of the equipment costs (Costeq):

Cost capital = G

∑ Cost

eq

(7)

9

where, G is the Lang factor, determined to be 1.6, which included 0.1 for electricity, 0.1 for instrumentation, 0.05 for lagging, 0.15 for civil work and 0.2 for installation. Equipment costs are usually correlated with the capacity factor using the following relationship: Cost eq = kYb

(8)

where k is the proportionality factor, Y is the capacity parameter and exponent b is typically within the range of 0.4-0.8, as demonstrated below (Brennan, 1998). In drying systems, the conveyors and heat exchangers are the main pieces of equipment. While the capacity factor of each of these is different, it is directly or indirectly dependent on the mass flow of air or steam. Equations 9 through 11 outline the equipment cost functions of individual items: belt dryer: Costeq = 2700Y Y is cross-sectional area heat exchanger: Costeq = 660Y0.7 Y is heat transfer area cover: Costeq = 1200Y0.5 Y is cover area

(9) (10) (11)

In calculating the cost of the belt dryer, the capacity parameter is affected by the belt cross-sectional area. This can be determined based on the fuel mass flowrate and the residence time of the wood. The mass flowrate (Mwood) was derived from the data in the previous section. The residence time (τwood) was considered to be the drying time of the pine wood, which was based on published date from both Fyhr and Rasmuson (1997) and Holmberg and Ahtila (2005). As the total amount of the wet fuel on the belt was known, calculated by: Mwood x (1+MC) x τwood)

(12)

where MC is the initial fuel moisture content, the required belt area can be estimated based on the unit area loading (Wload) of the wet fuel. Kiranoudis and Markatos (2000) recommended a maximum unit area loading of 50 kg/m2 on a wet basis. A unit area loading of 30 kg/m2, was used here and thus the effective area of drying (Aeff) was estimated by: A eff =

Mwood × (1 + MC) × τwood Wload

(13)

The equipment cost of the conveyor was then calculated according to Equation 13 and the cost function outlined in Equation 9. The cost of the cover was evaluated according to the cost function in Equation 11, based on the area that covers the conveyor belt. The length and width of the cover were slightly larger than the belt. The height of the cover was 6 m, which is commonly used in industrial belt dryer applications. For the heat exchanger, the heat transfer area (Aheat ex) was determined by: A heat ex =

Q h × (Tf − Twat )

(14)

where Q is the thermal flow rate, h is the heat transfer coefficient and Tf and Twat are the temperature of the flue gas and the water respectively. The heat exchanger was used to convert the hot water at 90°C into steam at temperatures of 140-180°C. During this process, the heat transfer rate consists of three stages: (i) the water temperature rises from 90°C to 100°C, (ii) the water evaporates to steam at 100°C, and (iii) the steam temperature increases to the desired temperature. Tf is the average value of the inlet and outlet flue gas temperatures for the heat exchanger. Because evaporation of liquid water to steam at 100°C is energy intensive, Twat was set to 100°C. The capital cost of the heat exchanger was then evaluated according to the cost function in Equation 10. Figure 8 shows the variation in capital costs with the final fuel moisture content at different flue gas temperatures. As expected, leaving the material at a higher final moisture level and/or using a higher operating temperature can reduce the capital costs. As the operation of the steam dryer is likely to cause corrosion problems, stainless steel can be partially used for equipment construction to minimise this issue. In the heat exchanger, for example, the tubes are constructed of stainless steel and the shell is constructed of carbon steel. Figure 9 shows the variation in capital costs with different final fuel moisture contents for 10

various steam conditions and equipment materials. The capital costs will be significantly increased if stainless steel is used.

Capital €) CapitalCost cost(millions (Million €)

3.0 90°C 90 °C

2.5

flue gas temperature

110 °C 110°C

2.0 1.5 1.0 0.5 0.0 0

0.2

0.4

0.6

Final Fuel Moisture Content (kg-water/kg-fuel)

CapitalCost cost (millions (Million €) Capital €)

Figure 8: Variation in the capital costs with final fuel moisture at an initial moisture content of 1.5 kg-water/kg-fuel for different flue gas temperatures. 10 9 8 7 6 5 4 3 2 1 0

140 °C C-steel 140°C carbon steel 160 °C C-steel 160°C carbon steel 140 °C SS316 140°C stainless 316 160 °C SS316 140°C stainless 316

0 0.2 0.4 0.6 Final Fuel Moisture Content (kg-water/kg-fuel)

Figure 9: in the capital costs with final fuel moisture at an initial moisture content of 1.5 kg-water/kg-fuel under different steam conditions and equipment materials.

As stated above, the overall costs of drying consist of both the capital and running costs. Running costs encompass all the costs associated with the operation of the dryer. The most significant are for the use of heat and electricity, as well as the maintenance costs, which are dependent on the annual operating time of the dryer and the price of energy. Maintenance costs are usually estimated as a percentage of direct capital costs; typically, values range from 2% to 11%, averaging around 5-6% (Brennan, 1998). Personnel costs and insurance are also often included in the running costs.

6. PROFITABILITY Based on the cumulative cash flow, the profitability was evaluated in terms of payback time, which is generally the main concern for investors. It is sometimes taken as the time from the commencement of the project to the recovery of the initial capital investment. When measuring profitability, the net present value (NPV) is used, which is a measure of the net cash benefit generated by a project and is utilized herein to evaluate the profitability of the designed drying processes. The NPV was calculated by: t =k

NPV =

C t − Cost main − Cost capital (1 + i) t t =0

∑

(15)

11

where t is an individual/specific year, k is the total number of years, Ct is the cash benefit in t years, Costmain is maintenance costs and i is the interest rate. The maintenance costs are generally around 5% of the capital costs. Expressions for calculating Costmain and Ct are as follows: Costmain = 0.05 Costcapital Ct = (Csave-Q•Cf) • τop

(16) (17)

where Csave is the saved fuel per MWh, Cf is the price of energy stored in the flue gas (€0.5/GWh here) and τop is the total number of operating hours in year t (in this case, 8400 hours). Since the heating rate for the water evaporation is Q (kJ/s), the total flue gas costs in this case would be Q•Cf. When the water content in the biomass is reduced, i.e. from an initial ratio of 1.5 down to 0.1 kg-water/kgfuel, the higher heating value (HHV) of the biomass will increase. This increased HHV will be beneficial in saving energy during the operation of the boiler to evaporate the same amount of water as removed from the fuel during drying. The saved fuel and thus the saved energy in the boiler (Csave) can be converted into a positive cash flow, as follows: Csave = Wevap × H latent × Cfuel

(18)

where Cfuel is the price of fuel. Wevap x Hlatent represents the total energy required to evaporate the desired amount of water in one hour, where Wevap can be found in Table 4. Cfuel generally depends on the type of fuel, time and other parameters but is considered here to be the same price as the biomass fuel used in the drying-boiler integrated process. The fuel price – how much the dried pine wood biomass can be sold for – is generally in the range of €6-20/MWh; here, €14/MWh was used for the calculations.

Net Present Value (millions €)

Figure 10 shows the variation in the NPV when the dryer is operating at a temperature of 90°C, using flue gas as the heat source – the first configuration described in Section 3. The finial moisture levels for the two cases considered here were 0.1 and 0.3 kg-water/kg-fuel. A return on investment should be achieved after 3 years of operation for the higher final moisture content and about 4 years for the lower moisture level. Figure 11 compares the 10-year NPV at different final fuel moisture levels. At an operating temperature of 110°C, a profit of €3.6m can be achieved after 10 years with the fuel moisture as low as 0.1 kg-water/kg-fuel. Increased fuel moisture contents lower the profit. At an operating temperature of 90°C however, the most profitable value is found a fuel moisture content of 0.3 kg-water/kg-fuel.

90°C, 0.1 kg-water/kg-fuel 90°C, 0.3 kg-water/kg-fuel

Duration of Operation (yrs)

Figure 10: Cumulative cash flow for biomass drying using flue gas at an operating temperature of 90°C and drying biomass from an initial moisture of 1.5 kg-water/kg-fuel to 0.1 and 0.3 kg-water/kg-fuel.

12

NPVValue (Million €) €) Net Present (millions

5.0 4.5

drying temperature

90 °C 90°C 110 °C 100°C

4.0 3.5 3.0 2.5 2.0 0

0.2

0.4

0.6

Final Fuel Moisture Content (kg-water/kg-fuel)

Figure 11: Variation in the net present value after 10 years with final moisture (initial moisture, 1.5 kg-water/kg-fuel).

Figure 12 shows the NPV for biomass drying using superheated steam – the second configuration described in Section 3 – at an operating temperature of 150°C; different steam recycle ratios are compared. As shown, 3-4 years of operation is expected to achieve a return on investment, as with the first configuration. Figure 13 shows the NPV after 10 years for different final fuel moisture contents at an operating temperature of 150°C; various steam recycle ratios were assessed. In general, the NPV decreases slowly as the final fuel moisture increases until it reaches levels of around 0.25 kg-water/kg-fuel, after which the profitability rapidly declines with increased final moisture. This indicates that drier biomass would yield a higher profits.

steam recycle ratio

Net Present Value (millions €)

0% 75%

Duration of Operation (yrs)

Figure 12: Cumulative cash flow for biomass drying using superheated steam at an operating temperature of 150°C and drying biomass from an initial moisture content of 1.5 kg-water/kg-fuel to 0.1 kg-water/kg-fuel. NPVValue (Million €) Net Present (millions €)

4.0 3.5 3.0 steam recycle ratio

2.5

R= 0% R= 50% R=75%

2.0 0 0.1 0.2 0.3 0.4 Final Fuel Moisture Content (kg-water/kg-solid)

Figure 13: Variation in the net present value after 10 years with final moisture, where the initial moisture is 1.5 kgwater/kg-fuel and the steam temperature is 150°C. 13

The results calculated and plotted in Figures 10 to 13 are all based on a fuel price of €14/MWh. As expected, profitability is very sensitive to the selling price of the fuel. Figure 14 outlines the effect of fuel price on the NPV after 10 years of operation. Selling the biomass fuel at a higher price obviously results in better profitability. Figure 14 also demonstrates that the fuel price needs to be greater than €8/MWh in order to see a return on investment after 10 years. Furthermore, the results indicate that drying fuels with a high moisture content will be beneficial.

NPV (Million €)

Net Present Value (millions €)

7 6

steam recycle ratio R=0%

5

R=75%

4 3 2 1 0 -1 0 -2

2

4

6

8

10 12 14 16 18 20

Fuel Price (€/MWh)

Figure 14: Variation in the net present value with fuel price after 10 years, for initial and final moisture contents of 1.5 and 0.1 kg-water/kg-fuel and a steam temperature of 150°C).

7. CONCLUSIONS This paper has studied the integration of a drying process into a power generation plant using two different forms of waste energy exiting from the process industry. The potential heating sources for biomass drying were both low-grade wastes – either the flue gas from the process or hot cooling water that could be used to form superheated steam. The dried biomass could then be provided as the fuel input for a subsequent 40 MW power station. A belt conveyor was the chosen dryer. According to the result herein, sufficient heat is contained in both the waste flue gases and the hot water exiting from the industrial process plant to be the heat source for biomass drying – in this case, white pine wood. The moisture levels can be reduced from 1.5 to 0.1-0.3 kg-water/kg-fuel, which is satisfactory for this to then be used as a fuel for combustion in the latter energy generation process, at a higher efficiency. By using flue gases as the heat source for drying, the capital costs would be in the region of €2.5 million. Although a higher flue gas temperature would reduce the capital costs, environmental issues may then become a problem, such as increased emissions. Using superheated steam as the drying medium however would mean that the capital costs would greater – about €3 million. To protect the equipment from corrosion, many components can be constructed from stainless steel, though this will double the equipment costs. In the selection of either the flue gas or superheated steam, the use of the flue gases would result in lower capital costs. Even though superheated steam is a good option in terms of short drying times, good heat recovery and environmental protection, the high capital costs associated with this dryer configuration is a considerable issue, particularly when stainless steel is used for some of the equipment components. Overall, for both the flue gas and steam drying configurations, 3-4 years of operation is expected to give a return on the initial investment at a fuel price of €14/MWh. However, profitability was found to be very sensitive to the biomass fuel selling price. It was calculated that this needs to be higher than €8/MWh to achieve a return on the investment after 10 years of operation.

ACKNOWLEDGEMENTS The authors would like to thank the UK Engineering and Physical Sciences Research Council (EPSRC Thermal Management of Industrial Process Consortium) and our industrial partners for their financial and technical support for this research programme. 14

NOMENCLATURE AND ABBREVIATIONS A Aeff Aheat ex b B C Cp,air Cp,water Cf Cfuel Csave Ct Costcapital Costeq Costmain Costrun G Gf Hf Hlatent HHV Hum h i k k L MC Mr Mwood NPV P Psat Q R t T T TDT Twat Wevap Wload Y Y

constant in Antoine equation effective area of drying heat transfer area of heat exchanger exponent constant in Antoine equation constant in Antoine equation specific heat of air specific heat of water price of flue gas price of fuel price of saved fuel cash benefit capital cost equipment cost maintenance costs running cost Lang factor mass flow rate of flue gas enthalpy of flue gas latent heat of water higher heating value humidity heat transfer coefficient interest rate total number of years proportionality factor longitudinal particle length moisture content molecular weight/mass dry mass flow of biomass net present value pressure saturated pressure thermal flow rate steam recycle ratio time temperature thickness total drying residence time Temperature of water evaporation rate of water unit loading of wood on the belt capacity parameter area

[-] [m2] [m2] [-] [-] [-] [kJ/kg-K] [kJ/kg-K] [€/MWh] [€/MWh] [€/MWh] [€] [€] [€] [€] [€] [-] [kg/s] [kJ/kg] [kJ/kg] [MJ/kg] [kg-water/kg-air] [W/m2K] [%] [yrs] [-] [m] [kg-water/kg-fuel] [-] [kg/s] [€] [mmHg] [mmHg] [W] [-] [s] or [yr] [K] or [°C] [m] [hrs] [K] [kg/s] [kg/m2] [-] [m2]

Greek symbols φ τ τop τwood

relative humidity drying time or operating time total operating hours in one year residence time of wood in the dryer

[-] [s or h/year] [hour] [s]

Subscripts air f in

air flue gas inlet, initial 15

out vapour water

outlet, final vapour water

REFERENCES Amos, W.A. (1998) Report on Biomass Drying Technology, NREL Contract No. DE-AC3683CH10093, National Renewable Energy Laboratory Barré, L. and Bilodeau, M. (1999) Drying residuals at low temperature with the Dry-Rex dryer, Pulp and Paper Canada 100, 132-138 Brennan, D. (1998) Process Industry Economics, UK: Institution of Chemical Engineers Bruce, D.M. and Sinclair, M.S. (1996) Thermal Drying of Wet Fuels: Opportunities and Technology, EPRI Report (TR-107109 4269-01) Chua, K.J., Mujumdar, A.S., Hawlader, M.N.A., Chou, S.K. and Ho, J.C. (2001) Batch drying of banana pieces – effect of stepwise change in drying air temperature on drying kinetics and product color, Food Research International 34, 721-731 Fredrikson, R.W. (1984) ‘Utilisation of wood waste as fuel for rotary and flash tube wood dryer operation’ Biomass Fuel Drying Conference Proceedings, Wisconsin, Office of Special Programs, University of Minnesota, 1-16 Fyhr, C. and Rasmuson, A. (1997) Some aspects of the modelling of wood chips drying in superheated steam, International Journal of Heat and Mass Transfer 40, 2825-2842 Gigler, J.K., van Loon, W.K.P., Vissers, M.M. and Bot, G.P.A. (2000) Forced convective drying of willow chips, Biomass and Bioenergy 19, 259-270 Ho, J.C., Chou, S.K., Mujumdar, A.S., Hawlader, M.N.A. and Chua. K.J. (2001) An optimisation framework for drying of heat sensitive products, Applied Thermal Engineering 21, 1779-1798 Holmberg, H. and Ahtila, P. (2005) ‘Optimization of the bark drying process in combined heat and power production of pulp and paper mill’, in Odilio, A.F., Eikevik, T.M. and Strommen, I. (eds.), Proceedings of the 3rd Nordic Drying Conference, Karlstad, Sweden Holmberg, H. (2007) Biofuel Drying as a Concept to Improve the Energy Efficiency of an Industrial CHP Plant, PhD Thesis, Helsinki University of Technology Huhtinen, M. and Hotta, A. (1999) ‘Combustion of bark’, in: Gullichsen, J. and Fogelholm, C.-J. (eds.), Chemical Pulping: Papermaking Science and Technology, 203-301 Kiranoudis, C.T., Maroulis, Z.B. and Marinos-Kouris, D. (1995) Heat and mass transfer model building in drying with multiresponse data, International Journal of Heat and Mass Transfer 38, 463-480 Kiranoudis, C.T. and Markatos, N.C. (2000) Pareto design of conveyor-belt dryers, Journal of Food Engineering 46, 145-155 Ogura, H, Yamamoto, T., Otsubo, Y., Ishida, H., Kage, H. and Mujumdar, A.S. (2005) A control strategy for chemical heat pump dryer, Drying Technology 23, 1189-1203 Spets, J.-P. and Ahtila, P. (2004) Reduction of organic emissions by using a multistage drying system for wood-based biomass, Drying Technology: An International Journal 22, 541-561 Tang, Z., Cenkowski, S. and Muir, W.E. (2004) Modelling the superheated-steam drying of a fixed bed of brewers’ spent grain, Biosystems Engineering 87, 67-77 Sheikholeslami, R. and Watkinson, A.P. (1992) Rate of evaporation of water into superheated steam and humidified air, International Journal of Heat and Mass Transfer 35, 1743-1751 van Deventer, H.C. (2004) Industrial Superheated Steam Drying, TNO report R 2004/239 Wimmerstedt, R. (1995) ‘Drying of peat and biofuels’, in: Mujumdar, A.S. (ed.) Handbook of Industrial Drying, New York: Marcel-Decker, 809-859 16