OBTAINING THE PROTOCOL OF ELIMINATION OF CARBONATES FROM MORROCAN OIL SHALE USING PLACKETT-BURMAN DESIGN

This work aimed to implement a screening experiment, to study the effects of different processing and formulation factors on the elimination of carbonates from Moroccan oil shale employing Plackett– Burman screening design. Different factors were proposed for study, such as temperature (°C), concentration (mol/l), ratio (s/l), processing time (h), mean particle size (mm), type of acid (CH3COOH, HCl), origin of row material (Tarfaya, Timahdit) and agitation (Yes, No). The quantity of carbonates eliminated was chosen as a response. The results revealed that concentration, ratio, processing time, mean particle size, type of acid and the origin of row material (Tarfaya, Timahdit), showed a significant effect on the quantity of carbonates eliminated, while the temperature and agitation had no effect. The predicted values were in agreement with the experimental values with a coefficient of determination (R 2 ) of 0.99. The model has been validated by experiments subsequent to optimized conditions.

This work aimed to implement a screening experiment, to study the effects of different processing and formulation factors on the elimination of carbonates from Moroccan oil shale employing Plackett-Burman screening design. Different factors were proposed for study, such as temperature (°C), concentration (mol/l), ratio (s/l), processing time (h), mean particle size (mm), type of acid (CH3COOH, HCl), origin of row material (Tarfaya, Timahdit) and agitation (Yes, No). The quantity of carbonates eliminated was chosen as a response. The results revealed that concentration, ratio, processing time, mean particle size, type of acid and the origin of row material (Tarfaya, Timahdit), showed a significant effect on the quantity of carbonates eliminated, while the temperature and agitation had no effect. The predicted values were in agreement with the experimental values with a coefficient of determination (R 2 ) of 0.99. The model has been validated by experiments subsequent to optimized conditions. Oil shale represents a significant potential resource, it is an interesting reserve in energy and their distribution in the word is more homogeneous than oil or gas natural. It is a real source of hydrocarbons, its interest lies in its content of kerogens, capable of turning into oil by increasing temperature and pressure (Qian et al., 2006) during its burial during geological time. Those benefits encourage Moroccan decision-makers to take more interest in its industrial exploitation but the presence of a mineral matrix intimately linked to the organic matters makes its exploitation not easy. This requires a concentration of organic matter by the elimination of carbonates which represent the majority of the raw material. The eight factors studied are the acid type, the temperature, the concentration, the ratio (l/s), the mean particle size (mm), the processing time, the agitation and the origin of row material.

Material and Methods:-Materials:-
The oil shale used in this work was collected from the R1 layer of Tarfaya and from Y layer of Timahdit, taken from South of Morocco. The shale was grinded (approximately between 0.5 and 1 mm in diameter). The organic matter of the oil shale was chemically linked to the mineral matter essentially formed by calcite, dolomite, silicate and clays. To free the rock from carbonate by dissolution in acids was carried out using HCl and CH 3

Decarbonatation:
Decarbonated oil shale was obtained by dissolution of carbonates using two different acids: hydrochloric acid (H) and acetic acid (A). The choice of acid is based on several tests performed in our laboratory and also according to previous studies. To optimize the decarbonatation conditions, various experiments were performed using different temperatures (20-50 °C), different concentrations (1-3 mol/l), ratios (L/S), mean particle size (0.5-1 mm), different processing times (4-24 h), agitation and origin of row material (R1-Y). 10 g of powdered rock (R1-Y) (mean particle size between 0.5 and 1mm) were prepared respecting the matrix of experiments in Table 3. After filtration, the solid residues were washed carefully with distilled water, dried at 70 °C and stocked in sealed plastic bags for future use. The used equation to calculate the amount of carbonates eliminated is given as follows.
where, Q m is the maximum quantity of carbonates eliminated,m i is the initial mass of oil shale used and m f is the final mass of oil shale obtained.

Plackett-Burman Design:
The Plackett Burman Design (PBD) is an efficient screening method to identify the important factors among large number of factors that influences a process. PBD was used to select the signi ficant factors out of eight factors considered in this study that influences the quantity of carbonates removed. For mathematical modeling, the following first-order polynomial model was used: Y = β 0 + Σβ i X i where, Y is the predicted response, β 0 is the model intercept and β i is the linear coefficient and X i is the level of the independent variable. Eight factors (5 continuous variables and 3 independent variables): temperature (°C) (A), concentration (mol/l) (B), ratio (s/l) (C), mean particle size (mm) (D), processing time (h) (E), origin of row material (F), type of acid (G), Agitation (H), has been studied to identify the significant decarbonatation factors of Tarfaya and Timahdit oil shales from morocco and the response (Y) is the amount of carbonates eliminated.

Statistical and data analysis:
Statistical analysis of the model was performed to evaluate the analysis of variance (ANOVA). Analysis includes Fisher test (F-test), its associated probability P (F) and the coefficient of determination (R 2 ) which

Matric of experience:
In this study, a 12-trial Plackett-Burman Design (PBD) was used to evaluate the eight factors. Each variable was evaluated at two levels: -1 for the low level and +1 for the high level. Table 2 represents the eight factors tested in Plackett-Burman Design and their levels. The experimental design of PBD (factors and tested range) is shown in Table 3.   According to the graph of the correlation, there is a distribution, the experimental performance is close to the theoretical line; the graph illustrates the good correlation between the values observed and that predicted with a coefficient of determination R 2 of the order of 1 (figure 1).
414 From table 4 it is observed that the value of R 2 = 0.99 and R 2 adjusted = 0.98 are very close. This is reflected in the fact that the observed variation is explained by the direct effects of the factors. This coefficient is very close to 1, so the quality of the adjustment of the Plackett-Burman Design chosen for the screening of the conditions is the best.

Test of analysis of variance anova:
In this analysis (Table 5), it should be noted that, the sum of the squares attributed to the total variation evaluated with 11 degrees of freedom is divided into a sum of two variations: one due to the regression which is estimated with 8 degrees of freedom, the other to the estimated residual variation with 3 degrees of freedom as defined in Table 5.
To assess the quality of the postulated model, Fisher Snedecor's test was used. On the basis of the comparison of the variance in the established model with respect to the variance of the residual, through the Fisher Snedecor test,we can say that for the model to be very significant at 95%, it is necessary that: F exp >> F (α, mod, ν res ), where α = 0.05 (5%).
The results of the analysis of the variance (Table 5) show that the experimental value of Scnedecor (F exp = 78.6479), which is the ratio between the square of the model and the mean square of the residue, is well above the value critical distribution (F (0.05; 8; 3) = 8.85) at a 95% confidence level at 8 and 3 degrees of freedom. Therefore, the ANOVA results given by the JMP 7 software is very significant with a confidence level of 95% and the model for response Y is considered compliant and of good quality. 415 Pareto diagram: The contributions of the factors are ranked in ascending order and then represented as a bar graph (Pareto diagram), shown in Figure 2. The Pareto chart review classifies the influence of different factors in the following order: the ratio (l/s), the type of acid, the concentration, the time, the mean particle size, the type of oil shale, the temperature, and then the agitation.

Prediction profiler for the response:
The prediction profile provided by the diagram below (figure 3) confirms that the factors that seem to be more influential are; concentration (B); ratio (s/l) (C); mean particle size (D); processing time (E); the origin of the row material (F), and the type of acid (G).
The mode of treatment of an analysis of the diagram also makes it possible to conclude that these factors affect the response in the antagonistic way. It is clear that an increase in concentration and ratio (s/l) leads to an increase in the ability to remove the carbonates from the Timahdit and Tarfaya oil shale.

Conclusion:-
The screening of Morocco's oil shale decarbonation conditions of Tarfaya (R1) and Timahdit (Y) was carried out by the experimental design methodology in the study of the effect of certain operating parameters, temperature (A); concentration (B); ratio (S/L) (C); mean particle size (D); processing time (E); origin of the raw material(F); acid type(G), and agitation (H).
The results obtained from eliminated carbonates led to the following conclusion: the most influential factors are concentration (B); ratio (S/L) (C); processing time (E); mean particle size (D); type of acid (G) and the origine of the row material (F); The optimization of the operating conditions allowed us to obtain a decarbanated material which has a decarbonation yield equal to 38.7%. The parameters were set at temperature = 37.67 °C, concentration = 1.93 (mol/l), ratio = 8.15 (s/l), mean particle size = 0.84 mm for 4.9 hour s with hydrochloride acid and Timahdit layer (Y) by providing agitation.