*3.1.3 Drying system*

Rice was dried in a laboratory drying equipment, especially designed and built for this purpose (Urumáquinas, Uruguay). It allowed controlling the drying air conditions including T, RH, and velocity with a precision of 0.6°C, 2.6%, and 0.02 m/s, respectively. The equipment also monitored the weight loss and grain temperature of the sample during drying. **Figure 3** shows a schematic of the drying system used.

The ambient air entered the system with the aid of a blower, which controlled the air velocity. A condenser and a vapor injector regulated the air humidity and resistors regulated the air temperature. A sensor of T and RH together with a sensor of velocity

#### **Figure 3.**

*Schematic of the drying equipment. 1-air entrance, 2-blower; 3-condenser, 4-resistors, 5-vapor injector, 6-velocity sensor, 7-air temperature and relative humidity sensor, 8-drying chamber, 9-load cell, and 10-PLC.*

was installed just before the drying chamber. Air conditions were set and controlled with the aid of a PLC (Secoin, Uruguay).

The rice sample was disposed in a tray with a perforated bottom to allow the air circulation. A temperature sensor was introduced in the rice sample to monitor the grain temperature. The sample weight was measured with the aid of a load cell, and the grain MC was calculated at different times using the initial MC and the weight loss:

$$\text{MCt} = 100 \times \left( 1 - \frac{IW}{Wt} \times \left( 1 - \frac{IMC}{100} \right) \right) \tag{1}$$

where MCt is the MC at a time t, IW is the initial weight of the sample, Wt is the weight of the sample at a time t, and IMC is the initial MC of the sample expressed on a wet basis.

All parameters (drying air T, RH, velocity, sample temperature, and weight) were registered every five minutes along each drying run.

#### *3.1.4 Rice drying*

Once the drying air reached the set condition, five hundred grams of paddy rice were put on the tray, arranged in a thin layer of one centimeter high, and introduced into the drying chamber.

A drying curve was built for each condition, leaving the rice to dry until no MC change was detected (at least ten consecutive measurements with grain MC differences among measurements lower than 0.5%). The drying curves were fitted to Page's Eq. (2):

$$\frac{\text{MC} - \text{EMC}}{\text{IMC} - \text{EMC}} = \exp(-k \ge t^n) \tag{2}$$

where MC is the moisture content at the drying duration t (h), EMC is the equilibrium moisture content, IMC is the initial moisture content, k is the drying rate constant (h�<sup>1</sup> ), and n is a dimensionless constant. MC, EMC, and IMC are expressed in a decimal dry basis.

Then, for each drying condition, rice was dried to a final MC of 17 � 0.7%, 15 � 0.7% and 13 � 0.7%. The time needed to reach each final MC at each drying air condition was calculated using the fitted Page's equation.

After drying, samples were submitted to one-hour tempering at the correspondent drying air temperature. Based on preliminary experiments, this was the minimum tempering time needed to minimize kernels' breakage after drying [15]. After tempering, samples with a final MC of 17% and 15% were dried in a chamber (Alfa-Laval Gruppe, Germany) at 20.5°C and 60% RH until a final MC of 13 � 0.5%. This gentle drying has a minimum impact on the grain quality, allowing to study the drying process at the MC range of interest. All experiments were performed in triplicate.

#### *3.1.5 Milling quality*

After drying and tempering, the samples were kept at room temperature for at least 72 hours. Then, the head rice yield (HRY) was determined.

Before milling, each sample was cleaned with a grain cleaner (Grainman, USA). Then, 100 g of clean paddy rice was hulled using a paddy husker (THU35B, Satake, *Improving the Efficiency of Rice Drying: Impact of Operational Variables on the Drying… DOI: http://dx.doi.org/10.5772/intechopen.112970*

Japan). The dehulled rice samples were milled with a laboratory rice polisher (TM05C, Satake, Japan) to a degree of milling (DOM) of 100 � 3, measured with a milling meter (MM1D, Satake, Japan). After milling, the broken kernels were separated using a trieur (Satake, Japan) and quantified using an image analyzer (Image 5, Selgron, Brazil). The results were expressed as grams of head rice obtained from 100 g of rough rice.

The head rice yield reduction (HRYR) during drying was defined as:

$$\text{HRYR} = \text{HRY}\_{\text{final}} - \text{HRY}\_{\text{initial}} \tag{3}$$

where HRYfinal was the HRY after the drying/tempering process at each drying condition evaluated, and HRYinitial represents the "maximum milling potential" of the rice lot. The HRYR is expressed in percentage points (pp), corresponding to the grams of milled head rice every 100 g of rough rice.

To determine the "maximum milling potential", four samples were gently dried in a chamber (Alfa-Laval Gruppe, Germany) at 20.5°C and 60% RH until a final MC of 13 � 0.5%. This air condition produces minimum fissuring and thus, minimal quality loss [10, 11]. Therefore, the HRYR obtained represents the maximum milling quality that can be achieved for the rice lot used.

#### *3.1.6 Statistical analysis*

The drying curves were fitted to Page's equation using the software JMP 12.0.

The standard deviation was calculated for the Page's equation parameters and the HRYR. Analysis of variance (ANOVA) was used to compare the n constant of Page's equation at different drying conditions. In the case of significative difference (p < 0.05), the Tukey test was applied to determine which are the values that differ.

The mean squared error (MSE) was calculated for Page's equation at each drying condition.

#### **3.2 Results and discussion**

**Table 2** shows the EMC calculated using the modified Chung-Pfost equation, the parameters k, n, and EMC from the fitted Page's equations, the corresponding mean squared error (MSE), and the drying durations to reach a final MC of 13%, 15%, and 17% (calculated using the fitted Page's equations) for each drying air condition.

The modified Chung-Pfost equation has been extensively used to calculate the EMC of grains. In Ref. [24], five different equations were compared and their suitability for describing the EMC of rough rice of different varieties (long and medium grain) was evaluated in a wide range of T and RH. The Chung-Pfost equation was the best model for describing equilibrium data.

Mathematical modeling of rough rice drying has also been studied by several researchers. A diffusion model assuming that liquid diffusion is the only moisture transfer inside the rice kernels has been used by some authors [25, 26]. However, solving this type of modeling is quite complex. Therefore, researchers usually use empirical or semiempirical models to simulate rice drying [27]. In Ref. [28], ten different models for continuous and intermittent drying of thin-layer rough rice were compared. The authors found that the Midilli model showed the best results but another four of them, including Page's model, were also adequate in describing the experimental data.




*EMC obtained from Chung-Pfost equation, Page's equation parameters (EMC, k, and n), MSE, and time needed to reach the final MC.*

### *Improving the Efficiency of Rice Drying: Impact of Operational Variables on the Drying… DOI: http://dx.doi.org/10.5772/intechopen.112970*

In Ref. [27], the suitability of twelve empirical and semiempirical models in defining thin-layer drying behavior of long-grain rough rice was studied. They also found that Midilli's model showed the best results, in part due to its high number of coefficients (four). The fact that it is a simplified version of the diffusion equation could also contribute, proving that liquid diffusion is the dominant transport mechanism in rough rice. Nevertheless, they found that Page's model was also suitable and was the most accurate among the two-parameter models. In agreement with these results, Pereira et al. [26] found that Page's model was the best model, among the six models studied, for describing continuous and intermittent drying of rough rice. Based on these previous reports, Page's model could be considered a simple (only two parameters), suitable model for describing thin-layer rough rice drying.

In the present study, except for the air condition at T = 35°C and RH = 50%, the EMC calculated using the modified Chung-Pfost equation was in quite good agreement with those obtained from Page's equation. In addition, Page's equations presented low MSE values (see **Table 2**), confirming its suitability for representing thin-layer drying of the long-grain rice variety Uy2.

**Table 2** also shows that n values are significantly different among some of the runs (p < 0.05). The k values can only be compared among those runs with n not significantly different. In those cases, k increased as the drying air temperature increased. Consequently, at constant EMC, the time needed to reach a certain grain MC decreased as temperature increased, as shown in **Table 2**. In agreement with our results, Chen et al. [2] found that higher temperatures may cause higher k values, even when the EMC was the same. This behavior could be expected since higher drying temperatures are associated with higher drying rates [16]. This is probably due to higher moisture effective diffusivities at higher drying temperatures [27, 29].

In Ref. [9], the authors found that drying above the Tg significantly increased the drying rate compared to drying below the Tg. This was attributed to the higher diffusivity observed in the rubbery state (compared to the glassy state). As previously exposed, in our experiments the drying constant (k), when comparable, also increased as the drying air T increased. In fact, when the drying air T increased from 35°C (below Tg) to 47°C (above Tg) at a constant EMC of 7%, the value of k increased almost 50%. However, when the drying air T increased from 47 to 55°C (both above Tg), the value of k only increased 10%. Therefore, the sharper increase of k in the former situation could be attributed, at least in part, to the glass transition phenomenon.

**Table 3** presents the HRYR of the rice dried at different drying air conditions and to different final MC. For rice taken to 17% and 15% MC, drying was finished in the drying chamber under mild conditions (T = 20.5°C and 62% RH). It can be observed that rice could be dried to a MC of 15% using drying air at temperatures as high as 47°C and RH as low as 27% maintaining a low HRYR (under 5 pp). Drying air temperatures of 55°C or higher increased drastically the HRYR.

During the last stage of the drying process (15–13% MC), milder drying air conditions should be applied to maintain a low HRYR. At 35°C, the HRYR was low at both RH tested. However, at 47°C the RH should be 57% or higher to keep the HRYR low.

This research brings important information on how the drying air conditions affect the drying rate and HRYR of a South American long-grain rice variety. The results could be used to implement a drying program that improves both aspects, using a more severe drying air condition until a MC of 15% to increase the drying rate, and then softening the drying air conditions at MC below 15%, to avoid an increase in the HRYR.


#### **Table 3.**

*Head rice yield reduction (HRYR) of rice dried to different moisture contents (MC: 17%, 15%, 13%) at different drying air conditions (T, RH). Rice with MC 17% and 15% was taken to a final MC of 13% under mild drying conditions (T = 20.5 °C/62% RH).*
