Mathematical Modeling of Drying Pattern of Ogi Produced From Two Types of Maize Grain

This paper presents thin layer modeling of ogi produced from yellow and white maize at varying soaking period and dried in the cabinet and oven at 50 oC. The moisture decrease for cabinet dried ogi produced from white maize from 49.0 11.5%, 49.5 to 11.32%, 46.5 to 12.33% and 46.12.29%. The drying rate for both oven and cabinet dried ogi produced from yellow maize decreased from 4.6 to 0.0525 kg/min, 4.5 to 0.0513 kg/min, 4.35 to 0.049 kg/min and 4.4 to 0.047 kg/min while for oven dried ogi followed a similar trend. The experimental data obtained were fitted to five thin layer models: Newton, Page, Herderson and Pabis, Two term and Wingh and Singh models. The values obtained for ogi produced from white maize and dried in the cabinet and oven at 50 oC for Newton model gave a lower R, χ, RMSE compared with respective values obtained from Page, Herderson and Pabis, two term, Wing and Singh models. The two terms model appear to be the best model among the five models used in this work and had higher R, lower χ, and RMSE. The ogi produced from yellow maize at varying soaking period of 24, 48, 72 and 96 hours and dried in cabinet dryer and fitted with two term showed model constants a, K0, b, K1 0.04315, 0.0388995, 0.919, 2.2 × 10 while the R, χ RMSE were 0.9933, 5.85 × 10 and 4.85 × 10 for ogi produced for 24 hours soaking, respectively. The soaking period does not seem to affect the moisture ratio and the thin layer drying model. However, the initial moisture and equipment seems to affect significantly.

Ogi produced from maize is the product obtained by fermentation of corn (Akingbala et al., 1981;Bolaji et al., 2011).The softened corn is washed and wet milled, sieved and sedement (Banigo & Muller, 1972).There has not been any substantial difference between the traditional and commercial manufacturing of ogi.Some modification have been introduced such as dry milling of maize into a fine flour and subsequent inoculation of the flourwater mixture with a culture of lactobacilli and yeast this is still not in general practice compared with the traditional method (Akingbala et al., 1981).According to Maskan 2001, the use high quality dried foods with good rehydration properties has become an interesting alternative to chemicals preservatives in food.This can be applied to ogi which is widely consumed within the geographical segments of Nigeria.
In view of the importance of ogi in the Nigerian diet, large scale production is appearing indispensable.The ogi material could be dried and package in polythene bags for an increased shelf life.There are many different types and variation of dryers, and selecting the proper dryer is crucial to achieving the desired results.Different types of dryers may be necessary depending on capacity, product quality, size, consistency, hours of operation, quantity of water to be evaporated, acidity of the product, operational environment and volatility of its flavor (S.Erenturk & K. Erenturk, 2007;Khazaei et al., 2008;Midilli, 2001).In addition, the nature of contact with food material, source of energy and the nature of the food material.
In most cases of application of thin layer drying, a finer grain obtained as result of millied process is rarely considered.Past work has been limited to vegetables, fruits, grains (legumes, cereals, seeds which particle size are quite larger compared with the particle size of fine ogi.Also, literature is very scanty on the application of thin layer to finely-grind product made through a sieve meshes (0.6 micron and below)-fine grains like ogi slurry obtained from fermented, milled, sieved, sedimented slurry consumed as weaning food and adult gruel.This work is attempted to find out if some common thin layer drying models may be relevant in describing the drying behavior of ogi.This may be subsequently useful in predicting, designing or selecting appropriate drying process and equipment for this product.

Method
Maize used for this experiment was obtained from a local market at Ketu, Lagos, Nigeria.Four kilogram were divided into four and each (1 kg) of the maize was weighed into bowl after thorough cleaning.Water was added and soaked for 24, 48, 72 and 96 hours.The soaked maize at different period was wet milled and were sieved with muslin clothe.The sediment was put in Muslim cloth and squeezed to drain the water ready for drying in two different dryer (cabinet and hot air oven) at 50 ºC.The ogi paste was spread to form a uniform thickness in the dryer stray (10 mm).This selection was necessary considering the gelatinization temperature of most starches and ogi (Bolaji et al., 2011).The moisture contents of paste were monitored at an interval of 10 minutes for 240 minutes (4 hours) subsequently at an interval of 2 hours for additional 4 hours.

Moisture Determination
The method of A.O.A. C (1990) was used to determine the moisture content of the ogi paste.A known mass of the paste was placed in an oven at 110 ºC for 3 hours, weighing was done.The final weight was taken when the product had cooled down inside a desiccator and the moisture content determined as a ratio of weight of water to weight of wet paste expressed in % as shown in Equation ( 1 Where MR

Results
The moist 46.5 to 12 white maiz The moisture ratio in Figure 1 showed that moisture in ogi samples decreased continually with drying time.The statiscal parameters of five common drying models used and fitted with RMSE and χ 2 used to evaluate the best model were presented in Tables 2, 3, 4, 5 and 6.The drying constant k, obtained for Newton model for ogi produced from maize soaked for 24 hours and dried in a cabinet dryer were 3.06 × 10 -3 with R 2 of 0.8963, χ 2 of 0.7071 × 10 -4 and RMSE of 0.021.Also for ogi produced from maize soaked at 48 hours and dried in the cabinet at 50 ºC, k is 3.35 × 10 -3 with R 2 of 0.77773, χ 2 of 1.5 ×10 -3 , RMSE of 0.26.Ogi produced from maize Maize soaked for 72 and 96 hours of soaking followed similar trend as shown in Table 2.The result for Oven dried Ogi from yellow maize at 50 ºC and 24 hours soaking time, revealed a k value to be 3.06 ×10 -3 , R 2 is 0.795, χ 2 is 3.167 ×10 -3 , RMSE is 0.054; and k, at 48 hours soaking time, is 4.45 × 10 -3 with R 2 of 0.882, χ 2 of 2.28×10 -3 , RMSE of 0.044 for ogi produced from yellow maize.The values obtained for ogi produced from white maize and dried in the cabinet and oven at 50 ºC, respectively revealed a similar trend.The R 2 were lower, compared with respective values obtained in page, Herderson and Pabis, Wang ang Singh models and two term models.The χ 2 and RMSE were contrary, their values were higher.However, the Two term odel appears to be the best model among the five models used in this work with higher R 2 , and lower χ 2, and RMSE.The ogi produced from yellow maize at varying soaking period of 24 hours and dried with cabinet dryer and fitted with two terms is as shown in Table 5.The result revealed constants a, K 0 , b, K 1 were 0.00594, 0.0202, 0.941, 0.00228, and R 2 , was 0.9901, χ 2 , 7.84 × 10 -4 , RMSE, 8.22 ×10 -5 , respectively.The values for ogi produced from 48 hours were 0.0103, 0.0125, 0.891, 0.00195 while the R 2 , χ 2 and, RMSE were 0.996823, 2.57 × 10 -4 and 2.85 × 10 -5 , respectively.Ogi produced from white maize and dried at the oven follow a similar trend.The R 2 , χ 2 and RMSE obtained for all the models for ogi produced from white maize and dried in the cabinet at the temperature used in this research work were lower as revealed by Newton, Page, Herdeson and Pabis and Two term models.The values of R 2 and other statistical parameters were lower compared to the findings of several previous works in fitting the model to the experimental data and only the values obtained from Two term model falls within range obtained for drying of apple (R 2 = 0.99869, χ 2 = 2.68 × 10 -4 ) and pumpkin (R 2 = 0.98952 and χ 2 = 2.31×10-3) as reported by Akpinar (2006); green table olives (R 2 = 0.9890 to 0.9987, RMSE = 0.009341 to 0.025469, and χ 2 = 8.9 x10 -5 to 6.54 ×10 -4 ) as reported by Demir et al. (2007); drying of figs (R 2 = 0.9912, χ 2 = 7.06 ×10 -3 , and RMSE = 0.074918) as reported by Doymaz (2005); black grapes (R 2 = (0.9794 to 0.9989, χ2 = 1.01x10 -4 to 1.772 ×10 -3 ) as reported by Doymaz (2006); prickly pear fruit (R 2 = 0.9993 and χ 2 = 1.1457×10 -4 ), drying shelled pistachios (R 2 = 0.9668, χ 2 = 4.756 ×10 -4 ) and unshelled (R 2 = 0.970 and χ 2 = 4.737×10 -4 ); natural solar drying of shelled pistachios (R 2 = 0.9380, χ 2 = 4.521×10 -4 ) and unshelled pistachios (R 2 = 0.9750 and χ 2 = 3.360×10 -4 ), according to Midilli and Kucuk (2003); drying of single apricot: (R 2 = 0.990, RMSE = 0.0487 and χ 2 = 0.002395) by Togrul and Pehlivan (2003); solar drying of sultana grapes: (R 2 = 0.973 and χ 2 = 0.005) as reported by Yaldiz et al. (2001).The study showed that drying ogi smaples in the cabinet dryer and hot air oven showed decrease in moisture content with increased drying at the same temperature of 50 ºC for 240 minutes.Drying of ogi may be achieved within the drying time employed in this research work however this is dependent on the quantity.Soaking period however did not reveal a significant effect on the drying pattern of ogi.The two term model showed a best model fit with Higher R 2 , lower χ 2 and RMSE followed by the Wang and Singh model.The values for R 2 , χ 2 and RMSE were highest in cabinet and oven dried ogi produced from yellow maize in all the models compared with the white ogi produced from white maize.The drying process was affected substantially by the consistent changing temperature gradient most expecailly at the at the beginignof the experiment.This was consistent with several researchers report (Akendo et al., 2008;Belghit et al., 2000;Falade & Abbo, 2007;Methakhup et al., 2005) of free water present at the start is vary, the rate of water removal is always higher during this at this stage (Guine´ et al., 2007).As the drying proceeds, the free water presents decreases quite rapidly, until the final stages when water was hardly available and the drying becomes very slow.This was reflected in the drying rate.Drying of ogi occurred predominantly in the falling rate period.This is indicative of the dominance of diffusion as physical mechanism governing moisture movements in the samples.This was consistent with the report by some researchers for greenbean (Rosello et al., 1997), okra (Gogus & Maskan, 1999), red chilli (Gupta et al., 2002), Carrot (Prabhanjan et al.,1995) and eggplant (Ertekin & Yaldiz, 2004;Falade et al., 2007).Generally, in this experiment, drying rates decreased with decreased moisture contents.Initially, there were higher drying rates when moisture contents were largest, after which, the drying rate decreased steadily with decreased moisture contents.This trend could be due to the removal of free moisture near the surface of the ogi paste at the early stages of drying.Moisture ratio decreased with increasing drying time.The solid content of ogi may be unconnected with the drying behavior (King, 1988;Lewicki, 2004;Mate et al., 1998;Falade et al., 2007).

Conclusion
The study has shown that drying pattern using the cabinet dryer and hot air oven decreased in moisture content with increased time of drying at the same temperature of 50 o C. Drying of ogi may be achieved within the drying time employed in this research work.However, the quantity and the maxium capacity of the drying equipment used determine will be important to determing the drying time.
The Two term model showed a best model fit with Higher R 2 , lower χ 2 and RMSE in this experiment.The cabinet dried ogi had Higher R 2 and lower χ 2 and RMSE compared with values obtained for oven dried ogi produced from yellow and white maize.This is an indication of the possible effect of the rate and efficiency of dehydration method. ).
Cabinet dried Ogi produced from White maize

Table 2 .
Newton models constants and statistical parameters

Table 3 .
Page models constants and statistical parameters

Table 4 .
Herderson and Pabis models constants and statistical parameters

Table 5 .
Two term models constants and statistical parameters

Table 6 .
Wang andSingh models constants and statistical parameters