MATERIALS AND METHODS

Briquetting process: Residual materials from palm oil mills, consisting of PKS and PF were collected from a mill in Bota, Perak. These loosely-bound raw materials were converted into solid fuel briquettes which has a relatively higher energy density compared to its raw form. Prior to compaction, the briquetting process involved drying, pulverizing and sieving. The oven dried raw materials were pulverized to particle sizes in the 63-500 μm range to establish some form of homogeneity. The process flow of the fuel briquetting process is shown in Fig. 2.

The compaction of the 40 mm diameter briquette was carried out in a 200 kN hydraulic press for every 10 g of powdered sample. The 200 kN force, corresponding to a 159 MPa pressing pressure, was within the range suggested by Yaman et al. (2000, 2001).

Fig. 2:	The briquetting process

Fig. 3:	Solid biomass fuel briquette of 10 g, 40 mm diameter

The diameter and height of the fuel briquette, as shown in Fig. 3 are 40 and 7 mm, respectively.

Characterization: The energy content of the fuel briquettes were measured using a LECO AC-350 bomb calorimeter. The proximate analysis for the samples was carried out in accordance to the Standard Test Method for Computational Analysis by Thermogravimetry. The ultimate analysis was carried out in a LECO CHNS 932 analyser which was calibrated with 51.78% carbon, 5.07% hydrogen, 20.13% nitrogen and 11.52% sulfur.

RESULTS AND DISCUSSION

Fuel properties: The HHV of PKS and PF were carried out for a total of three samples and their arithmetic average are used as outputs of this work. For the PKS briquettes, the average calorific value was found to be 19.38, 1.3 kJ g^-1 higher than the PF briquettes. Both materials had HHV comparable to lignite (brown coal) which is 16.28 kJ g^-1, as reported by Eugene and Theodore (1997). The heating values of the three materials are shown in Fig. 4.

The proximate and ultimate analyses of PKS and PF were conducted and the results are shown in Table 1 and 2, respectively.

The proximate analysis characterized the biomass materials by classifying the ingredients into three categories, namely volatile matter, fixed carbon and ash in percentage weight (dry wt. %). The higher fixed carbon content in PKS, for the experiment carried out in this work compared to PF, is expected and can be related to the heating values measured in the bomb calorimeter test. The results also indicate lower ash content in PKS compared to PF; higher ash values are detrimental to boiler operations as it contributes to the slagging of the boiler water tube surfaces. In practice, from an energy output point of view, a higher amount of PKS is preferred in the fuel mix entering the boiler furnace. A balance on the amount of PKS and PF in the fuel mix is required as the incineration of PKS on its own leads to incomplete burning and results in black smoke leaving the fuel stack.

Table 1:	Proximate analysis of palm kernel shell and palm fibre

Table 2:	Ultimate analysis of palm kernel shell and palm fibre

Fig. 4:	Calorific values of palm kernel shell (PKS) , palm fibre (PF) and lignite

The study of fuel characterization has in the past resulted in some useful correlations to predict parameters such as HHV. As a check with an establish model developed by Parikh et al. (2005), the characterization results in this work were used to estimate the fuel’s heating value. This value was then compared to the measured results from the bomb calorimeter tests. The comparison between the HHV results obtained via experiment and correlation is shown in Fig. 5.

The small difference in results is indicative of the consistency of the current test results compared to the database used to generate the HHV correlation. Results from the ultimate analysis showed that PKS had higher carbon compared to the PF samples. For a common fossil fuel such as coal, as the carbon content increases, so does the heating value. The consistency shown in the ultimate analysis results for PF and PKS with the earlier bomb calorimeter results is therefore accepted as a form of validation of the results presented in this work.

Optimization through mixing ratios: The bomb calorimeter test was used to find the experimental HHV of six palm oil mill waste material mixtures with their respective binder. The ratios were varied to find the best fuel based on energy content. Table 3 shows the mixing ratios used in the optimization activity. The six briquette material mixtures with binder could be grouped into two main categories; paper binder category and starch binder category. The S in the briquette named 60S:40F (paper) means Palm Kernel Shell (PKS), F means Palm Fibre (PF) and the p in parenthesis means paper as binder.

Table 3:	PKS and PF mixture with paper and starch binder

Fig. 5:	Comparison of high heating value (HHV) of palm kernel shell (PKS) and palm fibre (PF) found through experiment and correlation with the average absolute error of 3.74% indication

The binder material consisted of 10% of the total briquette weight, following the work by Husain et al. (2002) and Yunardi et al. (2011). The HHV results are shown in Figure 6. 60S:40F is the binderless briquette with the weight ratio of PKS to PF as 60 to 40.

Of all the six briquettes with binder, 60S:40F (paper) had the highest HHV. The 60S:40F briquette which had no binder material served as a reference in this study. Its HHV was 18.51, 0.12 kJ g^-1 lesser than the 60S:40F (paper) briquette. The proximate and ultimate analyses of mixing ratios 60S:40F, 60S:40F (paper) and 60S:40F (starch) were conducted and the results are shown in Table 4 and 5, respectively.

The HHV of 60S:40F (paper) briquette determined experimentally agreed with the HHV calculated using the correlation by Parikh et al. (2005), similar to the comparison with the briquettes made from pure PKS and PF. The HHV found experimentally was 18.63 kJ g^-1 whereas the one calculated using Parikh et al. (2005) correlation was 18.98 kJ g^-1, showing only a difference of 0.35 kJ g^-1. From the ultimate analysis, 60S:40F (paper) had the highest carbon content among all the materials mixtures. Taking the analogy to coal, the highest calorific value among the mixtures was due to its highest carbon content amongst them.

Table 4:	Proximate analysis of different 60S:40F briquettes

Table 5:	Ultimate analysis of different 60S:40F briquettes

Fig. 6:	High heating values of mixed-materials briquettes, s: Starch, p: Paper

Mechanical property of briquettes: Mechanical property is crucial for handling and storage of the biomass fuel briquettes. The effect of binder was justified from observation when 60S:40F, 60S:40F (starch) and 60S:40F (paper) briquettes were dropped from a height of 1-2 m, following the method adopted by Husain et al. (2002). Cracks resulted on the 60S:40F briquette was the most severe whereas 60S:40F (paper) was the least severe. The results are shown in Fig. 7.

Combustion and burning rate: The earlier HHV results had identified the mixing ratio of 60S:40F (paper) as the optimized solid fuel briquette mixing ratio. The combustion characteristics of the 60S:40F (paper) briquette was further studied by analyzing its burning profile. The briquette burning rate, representing its burning profile, was obtained via a thermogravimetry-weight loss method. The burning rate of the 60S:40F (paper) briquette is shown in Fig. 8.

The burning profile could be categorized into three regions: moisture release, rapid burning and slow burning region. These three sections were separated at around 163 and 426°C. Initially, as temperature increased, an obvious loss of sample mass occurred at 163°C. This loss represented the release of volatile matters and partial ignition. Rate of mass loss proceeded rapidly until the maximum value of 10.5 mg min^-1 was reached at 372°C. The mass loss rate, after the maximum, quickly slowed down to 1.6 mg min^-1 at 426°C.

Fig. 7(a-c):	Cracks resulted when (a) 60S:40F (b) 60S:40F (starch) and (c) 60S:40F (paper) briquettes were dropped from 1-2 m

Fig. 8:	Burning rate of 60S:40F (paper) briquette

The rapid mass loss was due to the combination from the release and then ignition of volatile matter which was as high as 71.81% on dry basis. The oily components within the PKS and PF samples are the likely reasons for the increase in the burning rate. Later into the burning process, the burning rate of the 60S:40F (paper) briquette was found to have slowed down due the partial coverage of ash over its surfaces. The oxygen diffusion through the ash layer had made it difficult for the remaining combustible sections of the briquette to be burned. The mass loss recorded during the slow burning process, up until 930°C, was less than 1 mg min^-1, a much slower process compared to the earlier expulsion of the volatile matter.

The combustion test was conducted in a combustion chamber to simulate the energy released by full sized 60S:40F (paper) fuel briquettes. The useful heat energy, Q_water, from the combustion of the briquettes was quantified through the temperature increase in the water jacket surrounding the combustion chamber (ΔT_water), mass flow of water (m_water) into the combustion chamber and the known specific heat of water (c_water). It was ideally assumed that the energy absorbed by the water, calculated by Q_water equal to m_water. c_water. ΔT_water, was equal to the amount of energy released by the combustion of the 60S:40F (paper) briquettes.

The rate of energy absorbed from the combustion chamber was calculated taking reference to the time taken for the water to reach its maximum temperature. In the experiment carried out, the amount of energy absorbed by 10.9 kg water in the water jacket and the rate of energy absorption were found to be 400.85 kJ and 0.1591 kJ sec^-1, respectively.

To explore the enhancements achieved through increasing the fuel surface area, the 60S:40F (paper) briquette design was modified to include a 4% increase in surface area via an indentation in the briquette centre. Similar combustion test procedures were carried out and found an increase of 13.3% in rate of energy released compared to the original briquette design. Results indicate potential enhancements in the fuel quality from the increase in exposed fuel surface area.

CONCLUSION

In this study, typical palm oil mill residues, combined in a mixture of 60% PKS and 40% PF with paper as its binding material, was found to be the optimum ratio for a viable solid fuel. A hole or a series of holes built into the briquette would enhance its combustion characteristics by increasing the exposed surface area. The study could be extended to determine the best briquette design that will further improve combustion characteristics.

ACKNOWLEDGMENTS

The authors would like to express their gratitude to Kilang Sawit Felcra Nasaruddin Km. 37 for providing the raw materials for the study and all technicians in Universiti Teknologi PETRONAS that are helpful in setting up the equipment for experiments.