- Email: [email protected]
An automated ash fusion test for characterisation of the behaviour of ashes from biomass and coal at elevated temperatures
Fuel 103 (2013) 454–466
Contents lists available at SciVerse ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
An automated ash fusion test for characterisation of the behaviour of ashes from biomass and coal at elevated temperatures Cheng Heng Pang a, Buddhika Hewakandamby a, Tao Wu b, Edward Lester a,⇑ a b
School of Chemical and Environmental Engineering, The University of Nottingham, University Park, NG7 2RD Nottingham, United Kingdom Division of Engineering, The University of Nottingham Ningbo China, 315100 Ningbo, China
h i g h l i g h t s " We show how images from the ash fusion test can generate a characteristic profile. " Each profile is linked to the ash composition of the pellet and source material. " These profiles help predict how the ash material might slag or foul in a boiler. " Biomass ash profiles are shown to be significantly different from coal profiles.
a r t i c l e
i n f o
Article history: Received 20 March 2012 Received in revised form 25 May 2012 Accepted 11 June 2012 Available online 17 August 2012 Keywords: Ash fusion test Characteristic temperature Dilatometry Sinter strength test Image analysis
a b s t r a c t The benefits of blending biomass with coal for power generation include less CO2 emissions and a reduced dependency on non-renewable fossil fuels. However, there is a need to understand the role of biomass during direct combustion and co-firing, particularly in terms of the effect of biomass use on ash slagging and fouling. A new image analysis based technique has been developed to characterise the behaviour of ashes from biomass, coal and coal/biomass blends using a single heating test at elevated temperatures. It is a reproducible test that combines the conventional ash fusion test, dilatometry and sinter strength test by means of image analysis. An oven is used to heat the cylindrical ash pellets from room temperature to 1520 °C, while the in-built camera captures still images of the samples throughout the temperature range. An automated image analysis code has been developed to provide behaviour profiles for each ash sample (across the temperature range) by quantifying dimensional changes upon heating. The error for the determined ash characteristic temperatures is approximately 15 °C, which is approximately 50% lower than for a conventional ash fusion test. Cylindrical ash samples from 9 biomasses (corn stover, DDG, DDGS, miscanthus, olive residue, wheat shorts, wheat, rapeseed, sunflower seed) and a standard UK coal (Daw Mill) were tested using this method and each was found to produce unique profiles. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction Whilst electricity generation from oil and natural gas generates little or no ash respectively [1], coal combustion can produce significant amount of ash that unavoidably leads to ash related problems. Ash slagging and fouling have always been major factors with regards to boiler design and operation, particularly as power stations now generally buy coals from all over the world. Coals from the world market have huge variations in total ash content and mineral composition [2]. Slagging deposits are normally located in the high temperature region of boilers with direct exposure to the combustion flame
⇑ Corresponding author. Tel.: +44 115 9514974. E-mail address: [email protected] (E. Lester). 0016-2361/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.fuel.2012.06.120
[3,4]. They are commonly found in the radiant section of boilers where the formation process is associated with the sintering and fusion of ash particles on surfaces, at temperatures in excess of 1000 °C [5]. Fouling deposits, however, occur in areas that are not directly exposed to thermal radiation, such as in the convection sections of a boiler [3,4]. This process involves lower temperatures, and is predominantly driven by the deposition of volatile inorganic species. Hurley and co-workers [6] reported that 20–40% of the ash produced in a boiler occur within the boiler itself, especially in the hoppers at the bottom section of the furnace and/or on the interior surfaces as slag. The remaining 60–80% exits with the flue gas for collection in mechanical and electrostatic separators. Although furnace design and operating conditions may contribute to the occurrence of slagging and fouling, ash characteristics play a major role. Ash chemical and mineral compositions determine its melting characteristics and fusion temperatures [7].
455
C.H. Pang et al. / Fuel 103 (2013) 454–466
A new challenge for boiler operators has been the introduction of biomass which has become increasingly popular in recent years as a carbon neutral energy source. Its lower melting temperature (compared to coal) can be problematic and limit the utilization of biomass in both direct combustion and co-firing. This phenomenon is mainly due to the high alkaline content of biomass ash, particularly in herbaceous biomasses [8]. Alkaline salts, such as alkaline silicates, melt and soften at low temperatures [9], forming a sticky liquid phase [10] which increases the stickiness of the surface of the ash deposit, leading to an increase in collection efficiency of incoming fly ashes and deposit growth [4,11]. These deposits are complex, heterogeneous, multiphase and porous materials [12] and commonly consist of alkali and alkaline earth metals [4], coupled with their chlorides, sulphates, carbonates and complex silicates [13]. Potassium is the main source of alkali in most biomass fuels [5,14] located in the inherent mineral matter and is the main cause of ash deposition and corrosion. Nielsen and co-workers [15] reported that potassium salts, mainly KCl or K2SO4, play a significant role in ash deposition by acting as glue bonding the individual fly ash particles together. Also, KCl is responsible for the corrosion of the superheater tubes in biomass fired boilers. However, KOH is formed when the level of chlorine is low. Other less stable potassium containing compounds may form but they are unlikely to enter the gas phase [16]. This is in contrast to coal, where sodium is the dominant and most problematic alkali metal [5,14]. The slagging and fouling propensity of a particular solid fuel can be predicted using various approaches including the ash fusion test [1,7,17–19] dilatometry/shrinkage [3,20,21], sinter strength test [17,22], viscosity measurements [7,23], various empirical indices [1,23–26] as well as pilot-scale trials. Each test has advantages and disadvantages which can include cost, reliability and complexity. The ash fusion test remains one of the most popular methods for studying fuels in terms of ash behaviour. The test is relatively inexpensive and simple with a significant number of publications that discuss its application [1,7,17–19] and methods to predict ash fusion temperatures [7,18,25,27–29]. Typically, pyramidal ash pellets are heated in a furnace under either oxidising or reducing conditions to over 1500 °C, and depending on the eventual shape and size of the pellets, four characteristic temperatures are determined for each sample (Fig. 1). The characteristic temperatures initial deformation, sphere, hemisphere and flow temperatures – indicate the behaviour of ashes in a boiler. The major drawback associated with the ash fusion test is the reliance on visual observation rather than an objective physical measurement i.e. the results are dependent on the subjective judgment of the operator. Inevitably there are inherent reproducibility issues, particularly between operators when a different operator is
Table 1 Types of samples used and their associated ash content. Fuel
Type
Ash content (wt.%)
Corn stover Wheat shorts Miscanthus Wheat Sunflower seed Rapeseed Olive residue Distillers dried grain (DDG) Distillers dried grain with solubles (DDGS) Daw Mill
Agricultural waste Agricultural waste Energy crop Industrial waste Industrial waste Industrial waste Industrial waste Industrial waste
4.2 4.6 2.8 3.3 1.9 4.9 13.4 3.9
Industrial waste
3.5
ECE/ISO classification 711 vitrinite reflectance (0.60%)
8.4
involved. Differences of 400 °C have been reported for the initial deformation temperature of a single sample obtained from different laboratories [3]. This work describes a new experimental test to characterise the complete behaviour of ashes by creating a fusion profile, rather than producing four specific temperature values. It is a test that combines ash fusion test, dilatometry and sinter strength test by means of image analysis. This measurement test is more reproducible because it is not based on manual identification. This test remains inexpensive (by using existing standard ash fusion kit) and straight forward. 2. Materials and methods A standard UK high volatile bituminous coal, Daw Mill as well as nine different biomasses were used for this study (Table 1). These biomasses cover a range including energy crops, agricultural wastes and industrial wastes. 2.1. Ashing Ash samples for this study were produced in a laboratory using a muffle furnace. All biomass samples were ashed at 650 °C. Large amounts of each biomass were ashed in order to generate the required amounts for pellet manufacture due to the inherent low ash contents (Table 1). The ashing temperature for coal is higher than that used for biomass. These temperatures represent the lowest temperatures to give loss on ignition (LOI) smaller than 5% after ashing. Preliminary results have shown that the difference in compositions between coal ashes prepared at 650 °C and 800 °C is
Fig. 1. Pyramidal ash pellet at different characteristic temperatures: (1): original pellet, (2): at deformation temperature, (3): at sphere temperature, (4): at hemisphere temperature, and (5): at flow temperature, where r is the radius of the pellet at hemisphere temperature (BS ISO 540:2008)[30].
456
C.H. Pang et al. / Fuel 103 (2013) 454–466
Fig. 2. Compression profile followed by the tensometer during the pelletization of cylindrical ash pellets.
within the acceptable experimental error of 10%. The ash fusion test was more sensitive to the presence of carbon (whilst measuring relative shape change during heating), than relatively small changes in elemental composition.
Fig. 3. Edge detection of ash pellets. (A) and (B) show the pellet at 300 °C and 845 °C respectively, while (A0 ) and (B0 ) are their detected edges.
2.2. Pelletization Ash samples were accurately weighed (1.00 ± 0.01 g for sinter strength test, 0.50 ± 0.01 g for the advanced ash fusion test) and then carefully loaded into a custom made die before being compressed using the Instron 3369 tensometer system to produce coherent cylindrical ash pellets. The exact same compression profile (Fig. 2) was followed for every sample where there was a 10 s holding time at the maximum compression pressure of 5000 psi. This resulted in a set of standard and very reproducible cylindrical pellets with 10 mm diameter and weighing exactly the same between pellets. 2.3. Advanced ash fusion test-picture analysis and graphing (PAnG) An SDAF2000d Ash Fusion Analyser was used for this test operating in an air atmosphere. The furnace was equipped with a Viewse VC-523D high resolution black and white closed circuit digital camera (CCDC). For each experiment, three different cylindrical ash pellets were simultaneously loaded and tested in the system. The furnace (loaded with samples) was heated up to 1520 °C at a rate of 10 °C/min, while the camera captured black and white images of the samples at a rate of 1 frame/°C increment. 2.4. Image analysis process The ash pellets were placed in the furnace in a row of three. The closed circuit digital camera (CCDC) was then focused on the samples through the viewing port. Matlab (version 2009a) was used to process each image, firstly to convert the RGB to 8 bit greyscale images. A greyscale image was then converted into Matlab matrix equivalent to the size of the image i.e. m n where m and n are the number of pixels in vertical and horizontal directions. This matrix has values from 0 (black) and 255 (white) depending on the shade of each pixel at each location. The method takes the initial image at ambient temperature to create a location template (horizontal and vertical) of each pellet. The surrounding furnace pixels are eliminated at this stage to create a pellet only image with a sufficient region around the pellet to allow movement and shape change tracking (should the pellet expand outside its original region). Each image was then passed through an edge detection loop to identify the boundary of the ash pellet against the base of the crucible. A gradient method was used to find the edge seeking the
Fig. 4. Characteristic profile of miscanthus ash produced from the PAnG test. The dotted lines separate the profile into different regions. Region I: unaffected by heat treatment; region II: sintering stage; region III: expanding stage; region IV: excessive melting/bubbling stage; ID: initial deformation temperature, F: flow temperature according to the conventional ash fusion test criteria.
change of grey level intensity at the edge [30]. Fig. 3 shows the detected edges at two different temperatures with the comparison to the digital image. Whilst the resolution of the image is relatively low (and this could be improved with a higher resolution camera), the position of the pellet is clearly identified. The contrast in each image decreases slowly with increasing temperature. With minor adjustments in image contrast across the full temperature range, the edges can be continuously detected. 2.5. Sinter strength test Cylindrical ash pellets were sintered and subsequently tested for strength. The sintering process was carried out at two different temperatures of 800 °C and 1100 °C using a muffle furnace. An average heating rate of 15 °C/min was used in order to avoid unnecessary cracks due to thermal shocks within the pellet which could affect its strength significantly. At increments of 150 °C a short hold time of 30 min was used to avoid damage to the pellets. An infrared thermocouple was used to check the maximum temperature of the samples during the five hours hold time. The strength of sintered ash pellets was measured using an automated Instron 3369 tensometer system. Each pellet was crushed with a load moving at a speed of 2 mm/min until the point of fracture. The failure point is the peak on the load-extension profile
C.H. Pang et al. / Fuel 103 (2013) 454–466
457
Fig. 5. Characteristic profiles (obtained from the PAnG test) of every sample tested together with their associated images (external structure) and mean strength at 800 °C and 1100 °C (obtained from the sinter strength test). X showing the void fractions (obtained from image analysis) indicated on the secondary y-axis, the dotted lines indicate the temperatures at which sintering was carried out for the sinter strength test in (E) showing that a particular pellet height occurs more than just once throughout the temperature range tested; showing the Sintering Point; (A) coal ash; (B) olive residue ash; (C) rapeseed ash; (D) sunflower seed ash; (E) miscanthus ash; (F) corn ash; (G) wheat ash; (H) DDGS ash; (I) DDG ash; (J) shorts ash.
(data recorded every 100 ms by the system) prior to an abrupt load drop of greater than 10% during crushing. The maximum load at the point of failure, together with the cross sectional area of the pellet
(measured before crushing using a vernier calliper) yielded its respective strength. Three pellets were manufactured and tested for each sample and the average taken as the sintered strength.
458
C.H. Pang et al. / Fuel 103 (2013) 454–466
Fig. 5 (continued)
3. Results and discussions 3.1. Advanced ash fusion test-picture analysis and graphing (PAnG) Compiling each image against temperature allows a shape/temperature profile to be generated. Fig. 4 shows a typical plot of relative height (pellet height at any given temperature/initial height
at ambient temperature) against temperature. This parameter was selected as the simplest means of describing shape change. Fig. 5 shows the profiles for each of the biomass samples and the coal. The key difference between samples is the temperature at which the different fusion/melting regions occur and the duration to which they hold. The samples showed lower fusion/melting regions than the coal, which is why the biomass profiles appear as
C.H. Pang et al. / Fuel 103 (2013) 454–466
459
Fig. 5 (continued)
though they have been shifted towards the lower temperature region. This difference in behaviour reflects the change in biomass elemental composition, compared to coal ash. This is considered more fully in Section 3.3.5. Fig. 6 shows cross section SEM images of ash pellets sintered at 800, 900, 1000 and 1100 °C (except for DDGS, DDG and shorts since they were completely melted at those temperatures) with the void percentage assessed by means of image analysis (Image J software). The void percentage is also presented in Fig. 5 as the secondary Y-axis (except for DDGS, DDG and shorts). 3.2. Sinter strength of the pellets Photographs of the sintered pellets are also shown in Fig. 5. Gibb’s work [22] on 31 coal ashes, shows increasing strength with increasing sinter temperature. Table 2 shows the average crushing strength results. Increasing the sintering temperatures from 800 °C to 1100 °C increases the strength of the pellet by a factor of 14. In contrast to Daw Mill, the biomass samples show decreasing strength with increasing temperature. By increasing the sintering temperature, miscanthus ash showed the largest decrease (by a factor of 8), while the strength of rapeseed and corn ash pellets was halved. Sunflower seed ash pellet strength decreased by a factor of three only olive residue ash pellets only showed a minimal decrease. The rest of the biomasses (DDG, DDGS, wheat and shorts) were excluded from the strength test as the pellets were completely melted and spread across the ceramic plate as thin films. All four biomasses were completely melted at 1100 °C, while DDG and shorts were completely melted even at 800 °C (Fig. 5G– J). The reasons for these effects are discussed in Section 3.3. 3.3. The link between ash fusion profile, pellet behaviour, strength and chemical composition Every biomass type appears to have its own characteristic profile (height/initial height against temperature plot). Four different types of behaviour appear to exist for the ash fusion profiles denoted by I, II, III and IV in Fig. 5, each representing a melting/fusion event that is taking place. Each biomass sample appears to show two or more of these regions. However, the temperature at which the melting/fusion events happen and the temperature range
across which they occur is specific to the biomass type, thus making each profile a finger print of an ash sample. Fig. 6 shows sectioned pellets prepared at 800, 900, 1000 and 1100 °C which help to explain the ash fusion profiles in Fig. 5. 3.3.1. Region I – unaffected by heat treatment This region is associated with the initial plateau at elevated temperatures which highlights the minimum temperature required before any physical changes in pellet dimensions occur. The profile for olive residue pellets shows a plateau up to 1200 °C which explains why pellets sintered at 800 °C (0.46 MPa) have a similar strength to those sintered at 1100 °C (0.43 MPa). It is apparent that the pellets were virtually unaffected by the sintering conditions which are both within region I (Fig. 5B). Daw Mill coal ash pellets gave similar strength value as olive residue (0.44 MPa) at 800 °C, where it is also within region I (Fig. 5A). This implies that no melting/fusion has taken place and that the value reflects the strength of an unsintered pellet. 3.3.2. Region II – sintering The fall in height is due to the shrinkage and densification of the pellet. Sintering causes the pellet to densify with an increase in grain size [5]. Before sintering (in region I), the ash pellet consists primarily of discrete particles that are isolated from neighbouring particles, but as the sintering process takes place, densification continues until the particles have substantial interconnection within the solid phase with bridges [12]. The cross-sectional SEM image in Fig. 6E of the miscanthus ash pellets of miscanthus in region II (800 °C and 900 °C) show decreased porosity and pore size due to the sintering process. Thus, it is evident that sintering alters the deposit microstructure significantly and simultaneously increases the solid fraction and density of the ash pellet/deposit. Sintering is commonly used to describe the particle-to-particle attachment during the heating process, which results in increased contact area between the ash particles [4] and causes the loosely attached particles to densify to a compact hard mass [31], thereby enhancing its strength. This effect is shown through the sinter strength test, where the pellet strength is much higher when sintered within the temperature range of region II as compared to other temperatures. Daw Mill coal ash pellet has higher strength
460
C.H. Pang et al. / Fuel 103 (2013) 454–466
Fig. 6. Miscanthus ash profile showing that a particular pellet height occurs more than just once throughout the temperature range tested.
when sintered at 1100 °C (region II) as compared to at 800 °C (region I), while corn, miscanthus and rapeseed ash pellets have much greater strengths when sintered at 800 °C (region II) as compared to at 1100 °C (region III). The only exception appears to be sunflower, which does not follow the general trend discussed. After cooling from 1100 °C to room temperature, the outer layer of the pellet remains oily. The sunflower seed ash appears to remain in region II beyond 1520 °C (Fig. 5D). Fig. 6D confirms that the change in voidage is relatively low compared to the other biomass pellets. This phenomenon requires further investigation particularly around the type of
mineral matter that contributed towards the final ash material in the pellets. If deposition occurs whilst the ash is within region II one would expect a harder denser deposit that would be difficult to remove. The temperature where region II (sintering) starts can be referred to as the ‘sintering point’. 3.3.3. Region III – expanding phase This region is the section of the characteristic profile with a positive gradient after region II. The increase in pellet height is partly associated with the thermal expansion of ash particles and the
C.H. Pang et al. / Fuel 103 (2013) 454–466
461
Fig. 6 (continued)
trapped air, and partly with the evolution of gases during the decomposition of ash components [32,33]. The gas is retained inside the fluid ash pellet causing it to expand and increase in height. The cross-sectioned SEM images (Fig. 6E) of miscanthus in region III (1000 °C and 1100 °C) show increased porosity (void percentage of 40.2% and 54.1% at 1000 °C and 1100 °C compared to 30.6% and 21.7% at 800 °C and 900 °C, respectively) and increased pore size due to the thermal expansion of various components and the evolution of product gases. These pellets were inherently weaker due to the increased pore size and degree of porosity [34,35]. The influence of voids on pellet strength is seen through the sinter strength test, where corn, miscanthus and rapeseed ash pellets show a reduced strength
when sintered at 1100 °C during which they show increased relative height or region III behaviour. Consequently, due to the increased porosity and pore size, if any ash deposits were to form in the actual boilers at the temperatures that create region III behaviour, a more porous and lower density ash will be easier to remove, especially through soot blowing. The temperature where region III starts can be referred to as the ‘expanding point’. 3.3.4. Region IV – excessive melting/degassing This region can be distinguished as the final depression in the characteristic profile. It is commonly accompanied by excessive bubbling, seen as noise or oscillations in the plot. This may also
462
C.H. Pang et al. / Fuel 103 (2013) 454–466
Fig. 6 (continued)
occur, to a lesser degree, in region III. Materials within this temperature range are highly fluid and this is evident from the excessive bubbling. The overall decrease in height shows that most of the materials are melted to a degree where they are unable to support the pellet structurally. This can be seen in Fig. 5G–J where DDG, DDGS, wheat and shorts are melted and spread across the ceramic plate as thin films. DDGS and wheat are melted at temperatures as low as 1100 °C, while DDG and shorts are completely melted even at 800 °C. DDG and shorts are already within region IV at 800 °C while DDGS and wheat are within region IV at 1100 °C. The evolution of gases ceases when all the gas is released or the heating stops, and hence the only major directional force acting on the ash material is
gravity. The formation of a thin molten film of ash on the plate at sintering temperatures of 1100 °C, and even 800 °C, suggests that most of the materials are melted and the formation of the film is due to gravitational pull on a low viscosity melt. Ash materials that are undergoing Region 4 type behaviour near the furnace wall will most likely form flowing slags. The temperature at which region IV starts can be referred to as the ‘excessive melting point’. 3.3.5. Link Between sintering point and ash chemical composition Fig. 7 shows the relationship between ash elemental compositions and ash behaviour. The ash behaviour is indicated by the sintering point (Table 3), illustrating the first event of melting. This
463
C.H. Pang et al. / Fuel 103 (2013) 454–466
Fig. 6 (continued)
Table 2 Pellet strengths of samples sintered at two different sintering temperatures. Some samples were completely melted and unable to undergo crushing test, thus labelled as N/A. Ash sample
38–75 38–75 38–75 38–75 38–75 38–75 38–75 38–75 38–75 38–75
Mean pellet strength (MPa) Sintering temperature of 800 °C
Sintering temperature of 1100 °C
0.44 0.46 9.63 6.40 29.13 8.76 30.86 N/A N/A N/A
5.97 0.43 4.37 2.28 3.46 4.92 N/A N/A N/A N/A
1/Shrinking Temperature
Daw Mill coal Olive residue Rapeseed Sunflower seed Miscanthus Corn stover Wheat shorts DDGS DDG Wheat
Particle size (lm)
Fig. 7. The linear relationship between ash behaviour (1/shrinking temperature) and ash composition (Group 1/Group 2), R2 = 0.9053.
relationship is produced using data from the biomass and coal samples tested. The Group 1 metals are predominantly in the form of potassium and sodium, while Group 2 metals are essentially calcium and magnesium. As the ratio of Group 1/Group 2 or (potassium + sodium)/(calcium + magnesium) increases, the inverse of the shrinking temperature increases linearly. This trend shows the lowering of shrinking temperature and hence melting temperature
with the increase in Group 1/Group 2 ratio. This is because potassium is the main source of alkali in most biomass fuels [5,14] located in the inherent mineral matter and it is therefore the main cause of ash deposition and corrosion. This is in contrast to coal, where sodium is the dominant and most problematic alkali metal [5,14]. 3.4. Reproducibility of the ash fusion test One of the key issues with the conventional ash fusion test is that the basis of the test itself is an observation rather than direct measurement [18], therefore it is subjective and dependent on the judgment of the operator [20]. This unavoidably leads to inherent reproducibility problems especially when different operators and/or apparatus are involved. It has been reported that if the ash fusion temperatures are determined within the same laboratory using the same operator, procedure and equipment, then the reproducibility is within 30–40 °C, and this will increase to 50– 70 °C if different laboratories or apparatus are used [25]. The poor repeatability and reproducibility of the conventional ash fusion test is an important issue [3]. The PAnG advanced ash fusion test uses image analysis to objectively quantify measure the various dimensions of the pellets across a wide temperature range. Fig. 8 shows profiles for a repeat
464
C.H. Pang et al. / Fuel 103 (2013) 454–466
Table 3 Shrinking points of samples obtained from the advanced ash fusion test. The shrinking points obtained from Test 1–3 are within ± 15 °C error. Shrinking point (°C)
Test 1 Test 2 Test 3 Average
Daw Mill coal
Corn stover
DDG
DDGS
Olive residue
Wheat shorts
Rapeseed
Sunflower seed
Wheat
Miscanthus
918 922 920 920 ± 2
726 715 719 720 ± 6
596 586 588 590 ± 5
621 616 623 620 ± 4
1114 1103 1113 1110 ± 6
698 686 686 690 ± 7
797 782 790 790 ± 8
780 778 782 780 ± 2
652 648 650 650 ± 2
788 788 794 790 ± 3
Fig. 8. Profiles of wheat ash from two separate runs of PAnG test, indicating the error being ±15 °C. Dotted lines indicate the different melting regions.
dimensional change over a large temperature range. In the conventional ash fusion test, four characteristic temperatures are determined, corresponding to initial deformation, sphere, hemisphere and flow temperatures while they are heated up to over 1500 °C. Other than the four characteristic temperatures, no other information is given, especially the changes in ash behaviour from one characteristic temperature to the other. As for the advanced ash fusion test, clear and distinct regions can be seen on the profiles indicating the different melting/fusion events that are occurring. The code also gives the conventional initial deformation and flow temperatures, which could be useful in existing ash slagging/fouling prediction models although Fig. 4 shows that the initial deformation temperature occurs within region III, after sintering temperature. This agrees with other workers who reported that the initial deformation temperature is not the temperature at which melting begins [18] nor is it the first shrinkage event [3] as there is considerable melting occurring at initial deformation temperature [21]. Fig. 5E shows the profile for miscanthus having pellets of the same heights at 800 °C and 1100 °C. From manual observation it could be assumed that the pellet has not changed. However, from the SEM images in Fig. 6E it can be seen that the pore size and porosity has changed dramatically. The same situation applies to the sample at 900 °C and 1000 °C. This could explain why the conventional ash fusion test can indicate two ash samples to have the same ash fusion temperatures despite having very different melting behaviours [18]. 3.6. Pellet type and the PAnG advanced ash fusion test
Fig. 9. TGA profile of Dextrin in air, illustrating that this compound is completely or almost completely decomposed at about 550 °C.
test on the same sample obtained from two separate pellets of wheat. Although the degree of expansion is clearly different, the critical temperatures at which melting/fusion events, for Regions I–IV, are occurring with reasonable accuracy. In total 68 samples (including samples not used in this publication) were analysed and the variations between pellets (three of each type) was found to be ±15 °C. Table 3 summarises the shrinking temperatures obtained from three separate runs on the PAnG test for the 10 samples used in this work. This process is therefore twice as accurate as the conventional ash fusion test with reproducibility reported to be at least 30–40 °C [25].
Pyramidal ash pellets are the standard for the conventional ash fusion test. These ash pellets are made manually by mixing a certain amount of ash with the prerequisite amount of Dextrin binder [36] before casting in a standard mould. The exposing surface of the mixture in the mould is then pressed to allow the pellet to be formed. Care must be taken to avoid damaging the pellet or distorting its shape, especially during the pellet removal from the mould. Although according to the British Standards BS ISO 540:2008 [36], the tips and edges should be checked for any damages, but internal imperfections are not visible. The presence or absence of only small amounts of material at the tip can lead to huge differences in ash fusion temperatures (as large as 400 °C [3]). Whilst the use of Dextrin is necessary for pyramidal pellet manufacture, its presence during heating can result in the formation of voids at low temperatures whilst it decomposes (Fig. 9). Pellets can be prepared without the use of binders if sufficient pressure is used during pelletization. For the advanced ash fusion test, a fixed amount of ash (0.500 g) was compressed in a custom made die system using a tensometer at known pressure of 5000 psi. 3.7. Potential uses for the PAnG advanced ash fusion test
3.5. Profiling using the PAnG advanced ash fusion test The PAnG advanced ash fusion test gives a complete behaviour profile of the ash pellet by quantifying all changes in shape using image analysis. The result is a characteristic profile detailing every
It is recognised that ashes obtained from actual power plant will differ from the laboratory produced ashes, but this ashing protocol remains a practical alternative to large scale trials, especially for new fuels where operational data is scarce.
C.H. Pang et al. / Fuel 103 (2013) 454–466
Power plant ashes appear to produce less gas during the ash fusion test, presumably as a result of the high heating rates and higher temperatures experienced in the boiler prior to deposition. This might not be a significant issue for the advanced ash fusion test since it is measuring relative changes across temperatures where melting/fusion events are occurring rather than the specific magnitude of the changes. Every ash sample will undergo sintering, expanding and excessive melting/bubbling during heating, but the key difference is the temperature at which these melting events occur, i.e. their sintering point, expanding point and excessive melting point. These characteristic temperatures are unique to each sample and will determine how the sample behaves at different temperatures. For a given boiler, one biomass may work well but others may not. For example, increasing the sintering temperature from 800 °C to 1100 °C increases the sinter strength of coal ash pellets, since the coal ash moves from region I (unaffected) to region II (sintering) on the characteristic profile. With the same temperature change in biomass ash, the strength decreases due to the movement from region II (sintering) to region III (expanding) on the characteristic profile. However, whilst samples could be ‘ranked’ in terms of slagging/ fouling propensity, the geometry and operating conditions of the boiler will be key factors in determining deposition behaviour. Gasification, processes could operate at temperatures above the initial deformation temperature in order to agglomerate ash to improve bed permeability, whilst operating below the flow temperature to prevent excessive clinkering [26]. As for other processes such as pf combustion, lower flue gas temperatures are preferred around convective heating surfaces to prevent fouling [37]. The advanced ash fusion profiles can inform the process of understanding how a fuel will behave in a specific boiler e.g. the profiles could be integrated into boiler models to predict locations where excessive slagging and fouling could occur. 4. Conclusions (1) A new experimental method has been developed to characterise fuel in terms of ashing behaviour. It is an adaptation of the conventional ash fusion test, incorporating dilatometry and sinter strength test by means of image analysis. (2) Compared to the conventional ash fusion test, the PAnG advanced ash fusion test; a. gives more information in the form of a complete behaviour profile with a new set of characteristic temperatures, namely sintering point, expanding point and excessive melting/bubbling point, whilst still giving initial deformation and flow temperatures. b. appears to be more accurate and reproducible by using image analysis edge detection and cylindrical ash pellets with no binders. (3) The ash fusion profile, coupled with the operating conditions/boiler temperature profile, enables the prediction of slagging/fouling propensity. (4) This method can be used to check if a fuel might be problematic for particular boiler systems.
Acknowledgments The authors gratefully express gratitude to all parties which have contributed towards the success of this project, both financially and technically (in terms of advice and/or assistance) particularly the Engineering and Physical Sciences Research Council (EPSRC –EP/F0608821/1), Ministry of Science and Technology of China (2008DFA61600), Ningbo Bureau of Science and Technology
465
(2008B10048), Dave Waldron, Dave Clift, Hui Yi, Shi Hong, Li Cheng, Hui Luan, I. Shong, Wei, Alicia, Voon, Caitlyn, Yee Ping, Vickie, Eliza, Francine and Melanie.
References [1] Regan JW. Impact of coal characteristics on boiler design. In: coal technology ’82 5th international coal utilization exhibition and conference. 1982. [2] Smith KL. The structure and reaction processes of coal. The Plenum chemical engineering series, New York: Plenum Press. 1994; xiv: p. 471. [3] Gupta SK et al. Ash fusion temperatures and the transformations of coal ash particles to slag. Fuel Process Technol 1998;56(1–2):33–43. [4] Zbogar A et al. Zbogar Shedding of ash deposits. Prog Energy Combust Sci 2009;35(1):31–56. [5] Loo SV, Koppejan J. Handbook/sjaak. The handbook of biomass combustion and co-firing. Earthscan; 2008. p. 442. [6] Hurley JP et al. Strength development at low temperatures in coal ash deposits. Prog Energy Combust Sci 1998;24(6):513–21. [7] Lolja SA et al. Correlation between ash fusion temperatures and chemical composition in Albanian coal ashes. Fuel 2002;81(17):2257–61. [8] Fernández Llorente MJ, Carrasco García JE. Comparing methods for predicting the sintering of biomass ash in combustion. Fuel 2005;84(14–15):1893–900. [9] Brown RC. TPB1/TPB Thermochemical Processing of Biomass: Conversion into Fuels, Chemicals and Power. John Wiley and Sons; 2011. p. 330. [10] Peter M. 7 Energy production from biomass (Part 1): overview of biomass. Biores Technol 2002;83(1):37–46. [11] Aho M, Ferrer E. Importance of coal ash composition in protecting the boiler against chlorine deposition during combustion of chlorine-rich biomass. Fuel 2005;84(2–3):201–12. [12] Robinson AL et al. Experimental measurements of the thermal conductivity of ash deposits: Part 2. effects of sintering and deposit microstructure. Energy Fuels 2000;15(1):75–84. [13] Jenkins BM, Bakker RR, Wei JB. On the properties of washed straw. Biomass Bioenergy 1996;10(4):177–200. [14] Baxter LL et al. The behavior of inorganic material in biomass-fired power boilers: field and laboratory experiences. Fuel Process Technol 1998;54(1– 3):47–78. [15] Nielsen HP et al. Deposition of potassium salts on heat transfer surfaces in straw-fired boilers: a pilot-scale study. Fuel 2000;79(2):131–9. [16] Jones JM et al. J15 an investigation of the thermal and catalytic behaviour of potassium in biomass combustion. Proc Combust Inst 2007;31(2):1955–63. [17] Cumming IW, Joyce WI, Kyle JH. Advanced techniques for the assessment of slagging and fouling propensity in pulverized coal fired boiler plant. J Inst Energy 1985:169–75. [18] Van Dyk JC et al. Coal and coal ash characteristics to understand mineral transformations and slag formation. Fuel 2009;88(6):1057–63. [19] Wall TF et al. False deformation temperatures for ash fusibility associated with the conditions for ash preparation. Fuel 1999;78(9):1057–63. [20] Kahraman H, Reifenstein AP, Coin CDA. Correlation of ash behaviour in power stations using the improved ash fusion test. Fuel 1999;78(12):1463–71. [21] Wall TF et al. Coal ash fusion temperatures – new characterization techniques, and implications for slagging and fouling. Prog Energy Combust Sci 1998;24(4):345–53. [22] Gibb WH. The slagging and fouling characteristics of coals II. A sintering test for the determination of fouling propensity. Power Industry Research 1 1981; 269–278. [23] Quon DHH, Wang SSB, Chen TT. Viscosity measurements of slags from western Canadian coals. J Eng Gas Turbines Power 1985;107(3):803–6. [24] Dunnu G, Maier J, Scheffknecht G. Ash fusibility and compositional data of solid recovered fuels. Fuel 2010;89(7):1534–40. [25] Seggiani M. Empirical correlations of the ash fusion temperatures and temperature of critical viscosity for coal and biomass ashes. Fuel 1999;78(9):1121–5. [26] van Dyk JC et al. Chemical fractionation tests on South African coal sources to obtain species-specific information on ash fusion temperatures (AFT). Fuel 2005;84(14–15):1768–77. [27] Evgueni J. Prediction of coal ash fusion temperatures with the FACT thermodynamic computer package. Fuel 2002;81(13):1655–68. [28] Liu YP, Wu MG, Qian JX. Predicting coal ash fusion temperature based on its chemical composition using ACO-BP neural network. Thermochim Acta 2007;454(1):64–8. [29] Özbayog˘lu G, Evren Özbayog˘lu M. A new approach for the prediction of ash fusion temperatures: a case study using Turkish lignites. Fuel 2006;85(4):545–52. [30] Parker JR. Algorithms for image processing and computer vision, vol. xiii. New York: Wiley Computer Publ.; 1997. [31] Skrifvars B-J, Backman R, Hupa M. Characterization of the sintering tendency of ten biomass ashes in FBC conditions by a laboratory test and by phase equilibrium calculations. Fuel Process Technol 1998;56(1–2):55–67. [32] Ilic M et al. Mineralogy and microstructure of sintered lignite coal fly ash. Fuel 2003;82(3):331–6. [33] Lin K-L, Chiang K-Y, Lin D-F. Effect of heating temperature on the sintering characteristics of sewage sludge ash. J. Hazard Mater 2006;128(2–3):175–81.
466
C.H. Pang et al. / Fuel 103 (2013) 454–466
[34] Palchik V, Hatzor YH. The influence of porosity on tensile and compressive strength of porous chalks. Rock Mech Rock Eng 2004;37(4): 331–41. [35] Wachtman JB. Mechanical properties of ceramics, vol. xxii. New York: Wiley; 1996. p. 448.
[36] British Standards ISO 540. Hard coal and coke. Determination of ash fusibility. BSI standards; 2008. [37] Pollock WH, Goetz GJ, Park ED. Advancing the art of boiler design by combining operating experience and advanced coal evaluation techniques. Medium: X Size; 1983. p. 102–17.
Contents lists available at SciVerse ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
An automated ash fusion test for characterisation of the behaviour of ashes from biomass and coal at elevated temperatures Cheng Heng Pang a, Buddhika Hewakandamby a, Tao Wu b, Edward Lester a,⇑ a b
School of Chemical and Environmental Engineering, The University of Nottingham, University Park, NG7 2RD Nottingham, United Kingdom Division of Engineering, The University of Nottingham Ningbo China, 315100 Ningbo, China
h i g h l i g h t s " We show how images from the ash fusion test can generate a characteristic profile. " Each profile is linked to the ash composition of the pellet and source material. " These profiles help predict how the ash material might slag or foul in a boiler. " Biomass ash profiles are shown to be significantly different from coal profiles.
a r t i c l e
i n f o
Article history: Received 20 March 2012 Received in revised form 25 May 2012 Accepted 11 June 2012 Available online 17 August 2012 Keywords: Ash fusion test Characteristic temperature Dilatometry Sinter strength test Image analysis
a b s t r a c t The benefits of blending biomass with coal for power generation include less CO2 emissions and a reduced dependency on non-renewable fossil fuels. However, there is a need to understand the role of biomass during direct combustion and co-firing, particularly in terms of the effect of biomass use on ash slagging and fouling. A new image analysis based technique has been developed to characterise the behaviour of ashes from biomass, coal and coal/biomass blends using a single heating test at elevated temperatures. It is a reproducible test that combines the conventional ash fusion test, dilatometry and sinter strength test by means of image analysis. An oven is used to heat the cylindrical ash pellets from room temperature to 1520 °C, while the in-built camera captures still images of the samples throughout the temperature range. An automated image analysis code has been developed to provide behaviour profiles for each ash sample (across the temperature range) by quantifying dimensional changes upon heating. The error for the determined ash characteristic temperatures is approximately 15 °C, which is approximately 50% lower than for a conventional ash fusion test. Cylindrical ash samples from 9 biomasses (corn stover, DDG, DDGS, miscanthus, olive residue, wheat shorts, wheat, rapeseed, sunflower seed) and a standard UK coal (Daw Mill) were tested using this method and each was found to produce unique profiles. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction Whilst electricity generation from oil and natural gas generates little or no ash respectively [1], coal combustion can produce significant amount of ash that unavoidably leads to ash related problems. Ash slagging and fouling have always been major factors with regards to boiler design and operation, particularly as power stations now generally buy coals from all over the world. Coals from the world market have huge variations in total ash content and mineral composition [2]. Slagging deposits are normally located in the high temperature region of boilers with direct exposure to the combustion flame
⇑ Corresponding author. Tel.: +44 115 9514974. E-mail address: [email protected] (E. Lester). 0016-2361/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.fuel.2012.06.120
[3,4]. They are commonly found in the radiant section of boilers where the formation process is associated with the sintering and fusion of ash particles on surfaces, at temperatures in excess of 1000 °C [5]. Fouling deposits, however, occur in areas that are not directly exposed to thermal radiation, such as in the convection sections of a boiler [3,4]. This process involves lower temperatures, and is predominantly driven by the deposition of volatile inorganic species. Hurley and co-workers [6] reported that 20–40% of the ash produced in a boiler occur within the boiler itself, especially in the hoppers at the bottom section of the furnace and/or on the interior surfaces as slag. The remaining 60–80% exits with the flue gas for collection in mechanical and electrostatic separators. Although furnace design and operating conditions may contribute to the occurrence of slagging and fouling, ash characteristics play a major role. Ash chemical and mineral compositions determine its melting characteristics and fusion temperatures [7].
455
C.H. Pang et al. / Fuel 103 (2013) 454–466
A new challenge for boiler operators has been the introduction of biomass which has become increasingly popular in recent years as a carbon neutral energy source. Its lower melting temperature (compared to coal) can be problematic and limit the utilization of biomass in both direct combustion and co-firing. This phenomenon is mainly due to the high alkaline content of biomass ash, particularly in herbaceous biomasses [8]. Alkaline salts, such as alkaline silicates, melt and soften at low temperatures [9], forming a sticky liquid phase [10] which increases the stickiness of the surface of the ash deposit, leading to an increase in collection efficiency of incoming fly ashes and deposit growth [4,11]. These deposits are complex, heterogeneous, multiphase and porous materials [12] and commonly consist of alkali and alkaline earth metals [4], coupled with their chlorides, sulphates, carbonates and complex silicates [13]. Potassium is the main source of alkali in most biomass fuels [5,14] located in the inherent mineral matter and is the main cause of ash deposition and corrosion. Nielsen and co-workers [15] reported that potassium salts, mainly KCl or K2SO4, play a significant role in ash deposition by acting as glue bonding the individual fly ash particles together. Also, KCl is responsible for the corrosion of the superheater tubes in biomass fired boilers. However, KOH is formed when the level of chlorine is low. Other less stable potassium containing compounds may form but they are unlikely to enter the gas phase [16]. This is in contrast to coal, where sodium is the dominant and most problematic alkali metal [5,14]. The slagging and fouling propensity of a particular solid fuel can be predicted using various approaches including the ash fusion test [1,7,17–19] dilatometry/shrinkage [3,20,21], sinter strength test [17,22], viscosity measurements [7,23], various empirical indices [1,23–26] as well as pilot-scale trials. Each test has advantages and disadvantages which can include cost, reliability and complexity. The ash fusion test remains one of the most popular methods for studying fuels in terms of ash behaviour. The test is relatively inexpensive and simple with a significant number of publications that discuss its application [1,7,17–19] and methods to predict ash fusion temperatures [7,18,25,27–29]. Typically, pyramidal ash pellets are heated in a furnace under either oxidising or reducing conditions to over 1500 °C, and depending on the eventual shape and size of the pellets, four characteristic temperatures are determined for each sample (Fig. 1). The characteristic temperatures initial deformation, sphere, hemisphere and flow temperatures – indicate the behaviour of ashes in a boiler. The major drawback associated with the ash fusion test is the reliance on visual observation rather than an objective physical measurement i.e. the results are dependent on the subjective judgment of the operator. Inevitably there are inherent reproducibility issues, particularly between operators when a different operator is
Table 1 Types of samples used and their associated ash content. Fuel
Type
Ash content (wt.%)
Corn stover Wheat shorts Miscanthus Wheat Sunflower seed Rapeseed Olive residue Distillers dried grain (DDG) Distillers dried grain with solubles (DDGS) Daw Mill
Agricultural waste Agricultural waste Energy crop Industrial waste Industrial waste Industrial waste Industrial waste Industrial waste
4.2 4.6 2.8 3.3 1.9 4.9 13.4 3.9
Industrial waste
3.5
ECE/ISO classification 711 vitrinite reflectance (0.60%)
8.4
involved. Differences of 400 °C have been reported for the initial deformation temperature of a single sample obtained from different laboratories [3]. This work describes a new experimental test to characterise the complete behaviour of ashes by creating a fusion profile, rather than producing four specific temperature values. It is a test that combines ash fusion test, dilatometry and sinter strength test by means of image analysis. This measurement test is more reproducible because it is not based on manual identification. This test remains inexpensive (by using existing standard ash fusion kit) and straight forward. 2. Materials and methods A standard UK high volatile bituminous coal, Daw Mill as well as nine different biomasses were used for this study (Table 1). These biomasses cover a range including energy crops, agricultural wastes and industrial wastes. 2.1. Ashing Ash samples for this study were produced in a laboratory using a muffle furnace. All biomass samples were ashed at 650 °C. Large amounts of each biomass were ashed in order to generate the required amounts for pellet manufacture due to the inherent low ash contents (Table 1). The ashing temperature for coal is higher than that used for biomass. These temperatures represent the lowest temperatures to give loss on ignition (LOI) smaller than 5% after ashing. Preliminary results have shown that the difference in compositions between coal ashes prepared at 650 °C and 800 °C is
Fig. 1. Pyramidal ash pellet at different characteristic temperatures: (1): original pellet, (2): at deformation temperature, (3): at sphere temperature, (4): at hemisphere temperature, and (5): at flow temperature, where r is the radius of the pellet at hemisphere temperature (BS ISO 540:2008)[30].
456
C.H. Pang et al. / Fuel 103 (2013) 454–466
Fig. 2. Compression profile followed by the tensometer during the pelletization of cylindrical ash pellets.
within the acceptable experimental error of 10%. The ash fusion test was more sensitive to the presence of carbon (whilst measuring relative shape change during heating), than relatively small changes in elemental composition.
Fig. 3. Edge detection of ash pellets. (A) and (B) show the pellet at 300 °C and 845 °C respectively, while (A0 ) and (B0 ) are their detected edges.
2.2. Pelletization Ash samples were accurately weighed (1.00 ± 0.01 g for sinter strength test, 0.50 ± 0.01 g for the advanced ash fusion test) and then carefully loaded into a custom made die before being compressed using the Instron 3369 tensometer system to produce coherent cylindrical ash pellets. The exact same compression profile (Fig. 2) was followed for every sample where there was a 10 s holding time at the maximum compression pressure of 5000 psi. This resulted in a set of standard and very reproducible cylindrical pellets with 10 mm diameter and weighing exactly the same between pellets. 2.3. Advanced ash fusion test-picture analysis and graphing (PAnG) An SDAF2000d Ash Fusion Analyser was used for this test operating in an air atmosphere. The furnace was equipped with a Viewse VC-523D high resolution black and white closed circuit digital camera (CCDC). For each experiment, three different cylindrical ash pellets were simultaneously loaded and tested in the system. The furnace (loaded with samples) was heated up to 1520 °C at a rate of 10 °C/min, while the camera captured black and white images of the samples at a rate of 1 frame/°C increment. 2.4. Image analysis process The ash pellets were placed in the furnace in a row of three. The closed circuit digital camera (CCDC) was then focused on the samples through the viewing port. Matlab (version 2009a) was used to process each image, firstly to convert the RGB to 8 bit greyscale images. A greyscale image was then converted into Matlab matrix equivalent to the size of the image i.e. m n where m and n are the number of pixels in vertical and horizontal directions. This matrix has values from 0 (black) and 255 (white) depending on the shade of each pixel at each location. The method takes the initial image at ambient temperature to create a location template (horizontal and vertical) of each pellet. The surrounding furnace pixels are eliminated at this stage to create a pellet only image with a sufficient region around the pellet to allow movement and shape change tracking (should the pellet expand outside its original region). Each image was then passed through an edge detection loop to identify the boundary of the ash pellet against the base of the crucible. A gradient method was used to find the edge seeking the
Fig. 4. Characteristic profile of miscanthus ash produced from the PAnG test. The dotted lines separate the profile into different regions. Region I: unaffected by heat treatment; region II: sintering stage; region III: expanding stage; region IV: excessive melting/bubbling stage; ID: initial deformation temperature, F: flow temperature according to the conventional ash fusion test criteria.
change of grey level intensity at the edge [30]. Fig. 3 shows the detected edges at two different temperatures with the comparison to the digital image. Whilst the resolution of the image is relatively low (and this could be improved with a higher resolution camera), the position of the pellet is clearly identified. The contrast in each image decreases slowly with increasing temperature. With minor adjustments in image contrast across the full temperature range, the edges can be continuously detected. 2.5. Sinter strength test Cylindrical ash pellets were sintered and subsequently tested for strength. The sintering process was carried out at two different temperatures of 800 °C and 1100 °C using a muffle furnace. An average heating rate of 15 °C/min was used in order to avoid unnecessary cracks due to thermal shocks within the pellet which could affect its strength significantly. At increments of 150 °C a short hold time of 30 min was used to avoid damage to the pellets. An infrared thermocouple was used to check the maximum temperature of the samples during the five hours hold time. The strength of sintered ash pellets was measured using an automated Instron 3369 tensometer system. Each pellet was crushed with a load moving at a speed of 2 mm/min until the point of fracture. The failure point is the peak on the load-extension profile
C.H. Pang et al. / Fuel 103 (2013) 454–466
457
Fig. 5. Characteristic profiles (obtained from the PAnG test) of every sample tested together with their associated images (external structure) and mean strength at 800 °C and 1100 °C (obtained from the sinter strength test). X showing the void fractions (obtained from image analysis) indicated on the secondary y-axis, the dotted lines indicate the temperatures at which sintering was carried out for the sinter strength test in (E) showing that a particular pellet height occurs more than just once throughout the temperature range tested; showing the Sintering Point; (A) coal ash; (B) olive residue ash; (C) rapeseed ash; (D) sunflower seed ash; (E) miscanthus ash; (F) corn ash; (G) wheat ash; (H) DDGS ash; (I) DDG ash; (J) shorts ash.
(data recorded every 100 ms by the system) prior to an abrupt load drop of greater than 10% during crushing. The maximum load at the point of failure, together with the cross sectional area of the pellet
(measured before crushing using a vernier calliper) yielded its respective strength. Three pellets were manufactured and tested for each sample and the average taken as the sintered strength.
458
C.H. Pang et al. / Fuel 103 (2013) 454–466
Fig. 5 (continued)
3. Results and discussions 3.1. Advanced ash fusion test-picture analysis and graphing (PAnG) Compiling each image against temperature allows a shape/temperature profile to be generated. Fig. 4 shows a typical plot of relative height (pellet height at any given temperature/initial height
at ambient temperature) against temperature. This parameter was selected as the simplest means of describing shape change. Fig. 5 shows the profiles for each of the biomass samples and the coal. The key difference between samples is the temperature at which the different fusion/melting regions occur and the duration to which they hold. The samples showed lower fusion/melting regions than the coal, which is why the biomass profiles appear as
C.H. Pang et al. / Fuel 103 (2013) 454–466
459
Fig. 5 (continued)
though they have been shifted towards the lower temperature region. This difference in behaviour reflects the change in biomass elemental composition, compared to coal ash. This is considered more fully in Section 3.3.5. Fig. 6 shows cross section SEM images of ash pellets sintered at 800, 900, 1000 and 1100 °C (except for DDGS, DDG and shorts since they were completely melted at those temperatures) with the void percentage assessed by means of image analysis (Image J software). The void percentage is also presented in Fig. 5 as the secondary Y-axis (except for DDGS, DDG and shorts). 3.2. Sinter strength of the pellets Photographs of the sintered pellets are also shown in Fig. 5. Gibb’s work [22] on 31 coal ashes, shows increasing strength with increasing sinter temperature. Table 2 shows the average crushing strength results. Increasing the sintering temperatures from 800 °C to 1100 °C increases the strength of the pellet by a factor of 14. In contrast to Daw Mill, the biomass samples show decreasing strength with increasing temperature. By increasing the sintering temperature, miscanthus ash showed the largest decrease (by a factor of 8), while the strength of rapeseed and corn ash pellets was halved. Sunflower seed ash pellet strength decreased by a factor of three only olive residue ash pellets only showed a minimal decrease. The rest of the biomasses (DDG, DDGS, wheat and shorts) were excluded from the strength test as the pellets were completely melted and spread across the ceramic plate as thin films. All four biomasses were completely melted at 1100 °C, while DDG and shorts were completely melted even at 800 °C (Fig. 5G– J). The reasons for these effects are discussed in Section 3.3. 3.3. The link between ash fusion profile, pellet behaviour, strength and chemical composition Every biomass type appears to have its own characteristic profile (height/initial height against temperature plot). Four different types of behaviour appear to exist for the ash fusion profiles denoted by I, II, III and IV in Fig. 5, each representing a melting/fusion event that is taking place. Each biomass sample appears to show two or more of these regions. However, the temperature at which the melting/fusion events happen and the temperature range
across which they occur is specific to the biomass type, thus making each profile a finger print of an ash sample. Fig. 6 shows sectioned pellets prepared at 800, 900, 1000 and 1100 °C which help to explain the ash fusion profiles in Fig. 5. 3.3.1. Region I – unaffected by heat treatment This region is associated with the initial plateau at elevated temperatures which highlights the minimum temperature required before any physical changes in pellet dimensions occur. The profile for olive residue pellets shows a plateau up to 1200 °C which explains why pellets sintered at 800 °C (0.46 MPa) have a similar strength to those sintered at 1100 °C (0.43 MPa). It is apparent that the pellets were virtually unaffected by the sintering conditions which are both within region I (Fig. 5B). Daw Mill coal ash pellets gave similar strength value as olive residue (0.44 MPa) at 800 °C, where it is also within region I (Fig. 5A). This implies that no melting/fusion has taken place and that the value reflects the strength of an unsintered pellet. 3.3.2. Region II – sintering The fall in height is due to the shrinkage and densification of the pellet. Sintering causes the pellet to densify with an increase in grain size [5]. Before sintering (in region I), the ash pellet consists primarily of discrete particles that are isolated from neighbouring particles, but as the sintering process takes place, densification continues until the particles have substantial interconnection within the solid phase with bridges [12]. The cross-sectional SEM image in Fig. 6E of the miscanthus ash pellets of miscanthus in region II (800 °C and 900 °C) show decreased porosity and pore size due to the sintering process. Thus, it is evident that sintering alters the deposit microstructure significantly and simultaneously increases the solid fraction and density of the ash pellet/deposit. Sintering is commonly used to describe the particle-to-particle attachment during the heating process, which results in increased contact area between the ash particles [4] and causes the loosely attached particles to densify to a compact hard mass [31], thereby enhancing its strength. This effect is shown through the sinter strength test, where the pellet strength is much higher when sintered within the temperature range of region II as compared to other temperatures. Daw Mill coal ash pellet has higher strength
460
C.H. Pang et al. / Fuel 103 (2013) 454–466
Fig. 6. Miscanthus ash profile showing that a particular pellet height occurs more than just once throughout the temperature range tested.
when sintered at 1100 °C (region II) as compared to at 800 °C (region I), while corn, miscanthus and rapeseed ash pellets have much greater strengths when sintered at 800 °C (region II) as compared to at 1100 °C (region III). The only exception appears to be sunflower, which does not follow the general trend discussed. After cooling from 1100 °C to room temperature, the outer layer of the pellet remains oily. The sunflower seed ash appears to remain in region II beyond 1520 °C (Fig. 5D). Fig. 6D confirms that the change in voidage is relatively low compared to the other biomass pellets. This phenomenon requires further investigation particularly around the type of
mineral matter that contributed towards the final ash material in the pellets. If deposition occurs whilst the ash is within region II one would expect a harder denser deposit that would be difficult to remove. The temperature where region II (sintering) starts can be referred to as the ‘sintering point’. 3.3.3. Region III – expanding phase This region is the section of the characteristic profile with a positive gradient after region II. The increase in pellet height is partly associated with the thermal expansion of ash particles and the
C.H. Pang et al. / Fuel 103 (2013) 454–466
461
Fig. 6 (continued)
trapped air, and partly with the evolution of gases during the decomposition of ash components [32,33]. The gas is retained inside the fluid ash pellet causing it to expand and increase in height. The cross-sectioned SEM images (Fig. 6E) of miscanthus in region III (1000 °C and 1100 °C) show increased porosity (void percentage of 40.2% and 54.1% at 1000 °C and 1100 °C compared to 30.6% and 21.7% at 800 °C and 900 °C, respectively) and increased pore size due to the thermal expansion of various components and the evolution of product gases. These pellets were inherently weaker due to the increased pore size and degree of porosity [34,35]. The influence of voids on pellet strength is seen through the sinter strength test, where corn, miscanthus and rapeseed ash pellets show a reduced strength
when sintered at 1100 °C during which they show increased relative height or region III behaviour. Consequently, due to the increased porosity and pore size, if any ash deposits were to form in the actual boilers at the temperatures that create region III behaviour, a more porous and lower density ash will be easier to remove, especially through soot blowing. The temperature where region III starts can be referred to as the ‘expanding point’. 3.3.4. Region IV – excessive melting/degassing This region can be distinguished as the final depression in the characteristic profile. It is commonly accompanied by excessive bubbling, seen as noise or oscillations in the plot. This may also
462
C.H. Pang et al. / Fuel 103 (2013) 454–466
Fig. 6 (continued)
occur, to a lesser degree, in region III. Materials within this temperature range are highly fluid and this is evident from the excessive bubbling. The overall decrease in height shows that most of the materials are melted to a degree where they are unable to support the pellet structurally. This can be seen in Fig. 5G–J where DDG, DDGS, wheat and shorts are melted and spread across the ceramic plate as thin films. DDGS and wheat are melted at temperatures as low as 1100 °C, while DDG and shorts are completely melted even at 800 °C. DDG and shorts are already within region IV at 800 °C while DDGS and wheat are within region IV at 1100 °C. The evolution of gases ceases when all the gas is released or the heating stops, and hence the only major directional force acting on the ash material is
gravity. The formation of a thin molten film of ash on the plate at sintering temperatures of 1100 °C, and even 800 °C, suggests that most of the materials are melted and the formation of the film is due to gravitational pull on a low viscosity melt. Ash materials that are undergoing Region 4 type behaviour near the furnace wall will most likely form flowing slags. The temperature at which region IV starts can be referred to as the ‘excessive melting point’. 3.3.5. Link Between sintering point and ash chemical composition Fig. 7 shows the relationship between ash elemental compositions and ash behaviour. The ash behaviour is indicated by the sintering point (Table 3), illustrating the first event of melting. This
463
C.H. Pang et al. / Fuel 103 (2013) 454–466
Fig. 6 (continued)
Table 2 Pellet strengths of samples sintered at two different sintering temperatures. Some samples were completely melted and unable to undergo crushing test, thus labelled as N/A. Ash sample
38–75 38–75 38–75 38–75 38–75 38–75 38–75 38–75 38–75 38–75
Mean pellet strength (MPa) Sintering temperature of 800 °C
Sintering temperature of 1100 °C
0.44 0.46 9.63 6.40 29.13 8.76 30.86 N/A N/A N/A
5.97 0.43 4.37 2.28 3.46 4.92 N/A N/A N/A N/A
1/Shrinking Temperature
Daw Mill coal Olive residue Rapeseed Sunflower seed Miscanthus Corn stover Wheat shorts DDGS DDG Wheat
Particle size (lm)
Fig. 7. The linear relationship between ash behaviour (1/shrinking temperature) and ash composition (Group 1/Group 2), R2 = 0.9053.
relationship is produced using data from the biomass and coal samples tested. The Group 1 metals are predominantly in the form of potassium and sodium, while Group 2 metals are essentially calcium and magnesium. As the ratio of Group 1/Group 2 or (potassium + sodium)/(calcium + magnesium) increases, the inverse of the shrinking temperature increases linearly. This trend shows the lowering of shrinking temperature and hence melting temperature
with the increase in Group 1/Group 2 ratio. This is because potassium is the main source of alkali in most biomass fuels [5,14] located in the inherent mineral matter and it is therefore the main cause of ash deposition and corrosion. This is in contrast to coal, where sodium is the dominant and most problematic alkali metal [5,14]. 3.4. Reproducibility of the ash fusion test One of the key issues with the conventional ash fusion test is that the basis of the test itself is an observation rather than direct measurement [18], therefore it is subjective and dependent on the judgment of the operator [20]. This unavoidably leads to inherent reproducibility problems especially when different operators and/or apparatus are involved. It has been reported that if the ash fusion temperatures are determined within the same laboratory using the same operator, procedure and equipment, then the reproducibility is within 30–40 °C, and this will increase to 50– 70 °C if different laboratories or apparatus are used [25]. The poor repeatability and reproducibility of the conventional ash fusion test is an important issue [3]. The PAnG advanced ash fusion test uses image analysis to objectively quantify measure the various dimensions of the pellets across a wide temperature range. Fig. 8 shows profiles for a repeat
464
C.H. Pang et al. / Fuel 103 (2013) 454–466
Table 3 Shrinking points of samples obtained from the advanced ash fusion test. The shrinking points obtained from Test 1–3 are within ± 15 °C error. Shrinking point (°C)
Test 1 Test 2 Test 3 Average
Daw Mill coal
Corn stover
DDG
DDGS
Olive residue
Wheat shorts
Rapeseed
Sunflower seed
Wheat
Miscanthus
918 922 920 920 ± 2
726 715 719 720 ± 6
596 586 588 590 ± 5
621 616 623 620 ± 4
1114 1103 1113 1110 ± 6
698 686 686 690 ± 7
797 782 790 790 ± 8
780 778 782 780 ± 2
652 648 650 650 ± 2
788 788 794 790 ± 3
Fig. 8. Profiles of wheat ash from two separate runs of PAnG test, indicating the error being ±15 °C. Dotted lines indicate the different melting regions.
dimensional change over a large temperature range. In the conventional ash fusion test, four characteristic temperatures are determined, corresponding to initial deformation, sphere, hemisphere and flow temperatures while they are heated up to over 1500 °C. Other than the four characteristic temperatures, no other information is given, especially the changes in ash behaviour from one characteristic temperature to the other. As for the advanced ash fusion test, clear and distinct regions can be seen on the profiles indicating the different melting/fusion events that are occurring. The code also gives the conventional initial deformation and flow temperatures, which could be useful in existing ash slagging/fouling prediction models although Fig. 4 shows that the initial deformation temperature occurs within region III, after sintering temperature. This agrees with other workers who reported that the initial deformation temperature is not the temperature at which melting begins [18] nor is it the first shrinkage event [3] as there is considerable melting occurring at initial deformation temperature [21]. Fig. 5E shows the profile for miscanthus having pellets of the same heights at 800 °C and 1100 °C. From manual observation it could be assumed that the pellet has not changed. However, from the SEM images in Fig. 6E it can be seen that the pore size and porosity has changed dramatically. The same situation applies to the sample at 900 °C and 1000 °C. This could explain why the conventional ash fusion test can indicate two ash samples to have the same ash fusion temperatures despite having very different melting behaviours [18]. 3.6. Pellet type and the PAnG advanced ash fusion test
Fig. 9. TGA profile of Dextrin in air, illustrating that this compound is completely or almost completely decomposed at about 550 °C.
test on the same sample obtained from two separate pellets of wheat. Although the degree of expansion is clearly different, the critical temperatures at which melting/fusion events, for Regions I–IV, are occurring with reasonable accuracy. In total 68 samples (including samples not used in this publication) were analysed and the variations between pellets (three of each type) was found to be ±15 °C. Table 3 summarises the shrinking temperatures obtained from three separate runs on the PAnG test for the 10 samples used in this work. This process is therefore twice as accurate as the conventional ash fusion test with reproducibility reported to be at least 30–40 °C [25].
Pyramidal ash pellets are the standard for the conventional ash fusion test. These ash pellets are made manually by mixing a certain amount of ash with the prerequisite amount of Dextrin binder [36] before casting in a standard mould. The exposing surface of the mixture in the mould is then pressed to allow the pellet to be formed. Care must be taken to avoid damaging the pellet or distorting its shape, especially during the pellet removal from the mould. Although according to the British Standards BS ISO 540:2008 [36], the tips and edges should be checked for any damages, but internal imperfections are not visible. The presence or absence of only small amounts of material at the tip can lead to huge differences in ash fusion temperatures (as large as 400 °C [3]). Whilst the use of Dextrin is necessary for pyramidal pellet manufacture, its presence during heating can result in the formation of voids at low temperatures whilst it decomposes (Fig. 9). Pellets can be prepared without the use of binders if sufficient pressure is used during pelletization. For the advanced ash fusion test, a fixed amount of ash (0.500 g) was compressed in a custom made die system using a tensometer at known pressure of 5000 psi. 3.7. Potential uses for the PAnG advanced ash fusion test
3.5. Profiling using the PAnG advanced ash fusion test The PAnG advanced ash fusion test gives a complete behaviour profile of the ash pellet by quantifying all changes in shape using image analysis. The result is a characteristic profile detailing every
It is recognised that ashes obtained from actual power plant will differ from the laboratory produced ashes, but this ashing protocol remains a practical alternative to large scale trials, especially for new fuels where operational data is scarce.
C.H. Pang et al. / Fuel 103 (2013) 454–466
Power plant ashes appear to produce less gas during the ash fusion test, presumably as a result of the high heating rates and higher temperatures experienced in the boiler prior to deposition. This might not be a significant issue for the advanced ash fusion test since it is measuring relative changes across temperatures where melting/fusion events are occurring rather than the specific magnitude of the changes. Every ash sample will undergo sintering, expanding and excessive melting/bubbling during heating, but the key difference is the temperature at which these melting events occur, i.e. their sintering point, expanding point and excessive melting point. These characteristic temperatures are unique to each sample and will determine how the sample behaves at different temperatures. For a given boiler, one biomass may work well but others may not. For example, increasing the sintering temperature from 800 °C to 1100 °C increases the sinter strength of coal ash pellets, since the coal ash moves from region I (unaffected) to region II (sintering) on the characteristic profile. With the same temperature change in biomass ash, the strength decreases due to the movement from region II (sintering) to region III (expanding) on the characteristic profile. However, whilst samples could be ‘ranked’ in terms of slagging/ fouling propensity, the geometry and operating conditions of the boiler will be key factors in determining deposition behaviour. Gasification, processes could operate at temperatures above the initial deformation temperature in order to agglomerate ash to improve bed permeability, whilst operating below the flow temperature to prevent excessive clinkering [26]. As for other processes such as pf combustion, lower flue gas temperatures are preferred around convective heating surfaces to prevent fouling [37]. The advanced ash fusion profiles can inform the process of understanding how a fuel will behave in a specific boiler e.g. the profiles could be integrated into boiler models to predict locations where excessive slagging and fouling could occur. 4. Conclusions (1) A new experimental method has been developed to characterise fuel in terms of ashing behaviour. It is an adaptation of the conventional ash fusion test, incorporating dilatometry and sinter strength test by means of image analysis. (2) Compared to the conventional ash fusion test, the PAnG advanced ash fusion test; a. gives more information in the form of a complete behaviour profile with a new set of characteristic temperatures, namely sintering point, expanding point and excessive melting/bubbling point, whilst still giving initial deformation and flow temperatures. b. appears to be more accurate and reproducible by using image analysis edge detection and cylindrical ash pellets with no binders. (3) The ash fusion profile, coupled with the operating conditions/boiler temperature profile, enables the prediction of slagging/fouling propensity. (4) This method can be used to check if a fuel might be problematic for particular boiler systems.
Acknowledgments The authors gratefully express gratitude to all parties which have contributed towards the success of this project, both financially and technically (in terms of advice and/or assistance) particularly the Engineering and Physical Sciences Research Council (EPSRC –EP/F0608821/1), Ministry of Science and Technology of China (2008DFA61600), Ningbo Bureau of Science and Technology
465
(2008B10048), Dave Waldron, Dave Clift, Hui Yi, Shi Hong, Li Cheng, Hui Luan, I. Shong, Wei, Alicia, Voon, Caitlyn, Yee Ping, Vickie, Eliza, Francine and Melanie.
References [1] Regan JW. Impact of coal characteristics on boiler design. In: coal technology ’82 5th international coal utilization exhibition and conference. 1982. [2] Smith KL. The structure and reaction processes of coal. The Plenum chemical engineering series, New York: Plenum Press. 1994; xiv: p. 471. [3] Gupta SK et al. Ash fusion temperatures and the transformations of coal ash particles to slag. Fuel Process Technol 1998;56(1–2):33–43. [4] Zbogar A et al. Zbogar Shedding of ash deposits. Prog Energy Combust Sci 2009;35(1):31–56. [5] Loo SV, Koppejan J. Handbook/sjaak. The handbook of biomass combustion and co-firing. Earthscan; 2008. p. 442. [6] Hurley JP et al. Strength development at low temperatures in coal ash deposits. Prog Energy Combust Sci 1998;24(6):513–21. [7] Lolja SA et al. Correlation between ash fusion temperatures and chemical composition in Albanian coal ashes. Fuel 2002;81(17):2257–61. [8] Fernández Llorente MJ, Carrasco García JE. Comparing methods for predicting the sintering of biomass ash in combustion. Fuel 2005;84(14–15):1893–900. [9] Brown RC. TPB1/TPB Thermochemical Processing of Biomass: Conversion into Fuels, Chemicals and Power. John Wiley and Sons; 2011. p. 330. [10] Peter M. 7 Energy production from biomass (Part 1): overview of biomass. Biores Technol 2002;83(1):37–46. [11] Aho M, Ferrer E. Importance of coal ash composition in protecting the boiler against chlorine deposition during combustion of chlorine-rich biomass. Fuel 2005;84(2–3):201–12. [12] Robinson AL et al. Experimental measurements of the thermal conductivity of ash deposits: Part 2. effects of sintering and deposit microstructure. Energy Fuels 2000;15(1):75–84. [13] Jenkins BM, Bakker RR, Wei JB. On the properties of washed straw. Biomass Bioenergy 1996;10(4):177–200. [14] Baxter LL et al. The behavior of inorganic material in biomass-fired power boilers: field and laboratory experiences. Fuel Process Technol 1998;54(1– 3):47–78. [15] Nielsen HP et al. Deposition of potassium salts on heat transfer surfaces in straw-fired boilers: a pilot-scale study. Fuel 2000;79(2):131–9. [16] Jones JM et al. J15 an investigation of the thermal and catalytic behaviour of potassium in biomass combustion. Proc Combust Inst 2007;31(2):1955–63. [17] Cumming IW, Joyce WI, Kyle JH. Advanced techniques for the assessment of slagging and fouling propensity in pulverized coal fired boiler plant. J Inst Energy 1985:169–75. [18] Van Dyk JC et al. Coal and coal ash characteristics to understand mineral transformations and slag formation. Fuel 2009;88(6):1057–63. [19] Wall TF et al. False deformation temperatures for ash fusibility associated with the conditions for ash preparation. Fuel 1999;78(9):1057–63. [20] Kahraman H, Reifenstein AP, Coin CDA. Correlation of ash behaviour in power stations using the improved ash fusion test. Fuel 1999;78(12):1463–71. [21] Wall TF et al. Coal ash fusion temperatures – new characterization techniques, and implications for slagging and fouling. Prog Energy Combust Sci 1998;24(4):345–53. [22] Gibb WH. The slagging and fouling characteristics of coals II. A sintering test for the determination of fouling propensity. Power Industry Research 1 1981; 269–278. [23] Quon DHH, Wang SSB, Chen TT. Viscosity measurements of slags from western Canadian coals. J Eng Gas Turbines Power 1985;107(3):803–6. [24] Dunnu G, Maier J, Scheffknecht G. Ash fusibility and compositional data of solid recovered fuels. Fuel 2010;89(7):1534–40. [25] Seggiani M. Empirical correlations of the ash fusion temperatures and temperature of critical viscosity for coal and biomass ashes. Fuel 1999;78(9):1121–5. [26] van Dyk JC et al. Chemical fractionation tests on South African coal sources to obtain species-specific information on ash fusion temperatures (AFT). Fuel 2005;84(14–15):1768–77. [27] Evgueni J. Prediction of coal ash fusion temperatures with the FACT thermodynamic computer package. Fuel 2002;81(13):1655–68. [28] Liu YP, Wu MG, Qian JX. Predicting coal ash fusion temperature based on its chemical composition using ACO-BP neural network. Thermochim Acta 2007;454(1):64–8. [29] Özbayog˘lu G, Evren Özbayog˘lu M. A new approach for the prediction of ash fusion temperatures: a case study using Turkish lignites. Fuel 2006;85(4):545–52. [30] Parker JR. Algorithms for image processing and computer vision, vol. xiii. New York: Wiley Computer Publ.; 1997. [31] Skrifvars B-J, Backman R, Hupa M. Characterization of the sintering tendency of ten biomass ashes in FBC conditions by a laboratory test and by phase equilibrium calculations. Fuel Process Technol 1998;56(1–2):55–67. [32] Ilic M et al. Mineralogy and microstructure of sintered lignite coal fly ash. Fuel 2003;82(3):331–6. [33] Lin K-L, Chiang K-Y, Lin D-F. Effect of heating temperature on the sintering characteristics of sewage sludge ash. J. Hazard Mater 2006;128(2–3):175–81.
466
C.H. Pang et al. / Fuel 103 (2013) 454–466
[34] Palchik V, Hatzor YH. The influence of porosity on tensile and compressive strength of porous chalks. Rock Mech Rock Eng 2004;37(4): 331–41. [35] Wachtman JB. Mechanical properties of ceramics, vol. xxii. New York: Wiley; 1996. p. 448.
[36] British Standards ISO 540. Hard coal and coke. Determination of ash fusibility. BSI standards; 2008. [37] Pollock WH, Goetz GJ, Park ED. Advancing the art of boiler design by combining operating experience and advanced coal evaluation techniques. Medium: X Size; 1983. p. 102–17.