Optimization of palm methyl ester and its effect on fatty acid compositions and cetane number

2.1. Materials and equipment’s used

Refined Palm oil was purchased from Ruchi Soya Industries Ltd. Pure Methanol (CH₃OH 99 % Pure), Sodium hydroxide pellets (NaOH 85 % Pure) were purchased from the Sigma-Aldrich Chemicals Ltd. The experimental setup consists of thermometer, magnetic stirrer with temperature and speed controller, separating funnel, conical and glass beakers of Borosil make were used for transesterification of methyl esters.

2.2. Transesterification

Raw oil is heated and filtered with surgical cotton to remove water molecules and suspended traces. Purified raw oil of measured quantity is heated up to 40°C with the help of a magnetic stirrer (reactor). Methanol (CH₃OH) added in Sodium hydroxide pellets (NaOH) formed as methoxide solution of required quantity is mixed with the raw oil. Proper care is to be taken while preparing the methoxide solution. If any undissolved sodium hydroxide pellets in methanol will lead to formation of soap. The mixture is stirred continuously for a time period by maintaining the 55 °C temperature. The temperature of the mixture should not exceed 62 °C this may result in the escaping of methanol and lead to incomplete reaction. After the desired reaction time and temperature, the glycerin will form. The high-density glycerin will settle in the bottom of the separating funnel. The glycerin is removed, and methyl ester is washed thoroughly with distilled water to remove the acids present in it. The water wash is repeated until a clear separation of methyl ester and oil. Fig. 1 represents the different stages of transesterification process and Fig. 2 represents the mechanism of transesterification process. The finally obtained palm methyl ester (PME) is tested for different physicochemical properties. The maximum percentage of PME yield was calculated by the formula as represented in the Eq. (1):

For the better yield of methyl esters there are four parameters which influence the maximum yield are methanol to oil ratio, catalytic concentration, reaction temperature and time. Taguchi based optimization technique is applied in this transesterification process by varying the above parameters.

Fig. 1Transesterification flow chart

Fig. 2Transesterification mechanism

2.3. Characterization

The obtained methyl esters were tested for different properties by following international standards. Table 2 represents the different property, equipment’s used and their standards for palm methyl ester and diesel.

2.4. Design of experiments

There are many factors which influence the biodiesel yield out of many molar ratio (methanol to oil ratio), catalytic concentration, reaction time and temperature are of most significant. In this research an attempt was made to analyze these parameters to achieve the maximum biodiesel yield. Dr. Taguchi has introduced the statistical technique to analyze the different parameters which influence the process conditions. By implementing these techniques can reducing the experimental cost with minimum number of experiments. This technique uses orthogonal array for designing the experiments and it can predict the important influencing parameters for biodiesel yield. In the present analysis L9 orthogonal array is chosen for optimization of PME. Eq. (2) represents the number of experimentations that has to carried out and it is clear that number of experimental runs depend upon chosen level and number of factors. From the equation it is found that there are Nine set of experiments that has to conduct in order to calculate response parameters:

where: $L$ represents the chosen level, $F$ – the number of factors.

3.1. Fatty acid compositions

Fatty acid compositions play a vital role during the combustion in the combustion chamber. In general, these fatty acids are mainly divided into two major categories, one is saturated fatty acids and another is unsaturated fatty acids, unsaturated fatty acids further divided as mono and poly unsaturated fatty acids. The exhaust emissions like unburnt HC, NOx, smoke and etc. will depends on the fatty acid composition present in the biodiesel. If the presence of excess linoleic fatty acid in the biodiesel may lead to formation of NOx and exhibits low thermal efficiencies [12, 13]. And it also difficult to convert the raw oil into methyl ester if the presence of free fatty acids (FFA) more than 2.5 %. Table 3 represents the different fatty acid composition of PME and Fig. 3 represents the contribution of saturated and unsaturated fatty acids in PME.

Fig. 3Saturated and unsaturated contribution of PME

3.2. Prediction of Cetane number (CN) using fatty acid compositions (FAC)

The quality of combustion depends on the quality of the fuel, Cetane number (CN) is known for the diesel quality index. In general, the CN for biodiesel is always greater than the conventional diesel fuel. The cost involved in determining the CN through experimental approach is high and sometimes even less accurate due to the experimental errors. In order to overcome the experimental cost and errors different methods have been proposed to predict the CN. Van Gerpen [14] tested the CN of soybean biodiesel and concluded that CN is depends on the distribution of fatty acid composition (FAC) in the oil. The CN is more in saturated FAC, less in unsaturated FAC and moderate in mono unsaturated FAC [15]. A. I. Bamgboye [16] suggested that, with the FAC it is possible to predict the CN. Eq. (3) represent the relation between biodiesel FAC and CN. From the FAC of PME the predicted CN is 55.38 and by experimental approach [17] the CN is 54 [16]:

where $K$, $a$, $b$, $c$, $d$, $e$, $f$, $g$, $h$ are constants taken from regression analysis and $X_1$ to $X_8$ are the % of FAC:

where $X_2$ to $X_8$ were taken from the Table 2.

3.3. Optimization by Taguchi

The main objective of the present analysis is to maximize the PME yield, for this purpose Taguchi based optimization technique was adapted. Initially a nine set of experiments were performed, and yield percentages were calculated for each case by varying the four different parameters of molar ratio, catalyst concentration, reaction time and temperature. The obtained experimental results were furnished in Minitab software and analyzed for larger the best (Eq. (4)). Table 4 represents the different parameters which are varied at three levels. Table 5 represents the PME yield and signal to noise ratio level (SNRL) for nine set of experiments:

where:

Table 6 represents the catalyst concentration, molar ratio and reaction temperatures for different levels. Here the SNRL represents the algebraic mean of particular parameter at that level. The delta represents the difference in maximum and minimum values of particular parameters, based on the higher and lower values of delta, the ranks were assigned. Catalyst concentration is ranked as one and it is determined as influencing parameter for the PME yield.

3.4. Analysis of variance

The signal to noise ratio (SNR) represents the ratio of mean value of biodiesel yield to standard deviation. SNR helps in predicting the optimum level of parameters with respect to optimum conditions to maximize the biodiesel yield. The exact parameter which influence the biodiesel yield at that particular condition is unable to predict by using SNR. Hence, a statistical analysis of variation is used for the biodiesel yield (ANOVA). In the present analysis ANOVA test is used for the prediction of PME and also the different parameters that influence the PME yield. Table 7 represents the different process parameters for $F$ and $P$ values. The percentage of contribution was also calculated by the contribution factor formula as shown in the Eq. (5). Higher the $F$ value represents the significant parameter in the preparation of methyl esters and its yield. The Contribution factor will also revel the significant process parameter but this $F$ value give the reconfirmation. From the Table 5 the maximum $F$ value (28.86) and the corresponding low $P$ value (0.006) are significant:

where:

$S{S}_f$ represents the sum of the squares for the $f$th parameter, $S{S}_t$ represents the total sum of the squares of all parameters, $N$ is the number of experiments at level $J$ of factor $f$.

3.5. Signal to noise ratio (SNR)

Fig. 4, 5 represents variations of molar ratio, catalyst concentration, reaction time and temperature of the reaction in terms of signal to noise ratio. The maximum value in each parameter represents the optimum level for PME yield. Therefore, at molar ratio of 5:1, catalyst concentration of 4.0 grams, reaction time of 180 minutes and temperature of 44°C gives the maximum yield for PME.

Fig. 4SNR plot for molar ratio and catalyst concentration

Fig. 5SNR plot for reaction time and temperature

3.6. Regression equation

The optimum yield of PME is calculated by using the regression equation. From the SNR graphs the optimum value of each parameter for the maximum yield are obtained. By substituting these optimum values in the regression equation, the optimum yield value is 92.06 %.

Yield = 73.21 – 0.100 (Molar ratio) + 5.62 (Catalyst Concentration) + 0.0589 (Reaction time) – 0.376 (Reaction Temperature) = 92.06 %.

The optimum values from SNR plots of molar ratio (5:1), catalyst concentration Eq. (4), reaction time (180 min) and reaction temperature (44 °C) were considered for experimentation. By following these conditions, the experimentation is performed and for better accuracy the experimentation is repeated for three trails, the average of three experimental yield is 91.65 % as shown in the Table 8.