Response surface methodology for optimization of process and machine parameters for sieving of gram flour using a multi-purpose portable sieve shaker

Introduction

Gram flour (besan) is an important pulse-based flour widely used in food processing. After grinding, the material contains particles of different sizes, and effective sieving is required to obtain a relatively uniform product and remove coarse particles. Conventional manual sieving is labour-intensive, time-consuming and dependent on operator skill. The development of an affordable and portable mechanized sieve shaker is therefore useful for small-scale food-processing applications.

Sieving is an important post-grinding operation because it contributes to particle-size classification, removal of unwanted coarse material and improvement of product uniformity. The performance of a sieve depends on operating and material-related factors such as feed rate, sieve speed, mesh size, moisture and particle characteristics. Mechanized systems can provide more uniform motion and reproducible operating conditions than manual sieving. The present research work entitled "Development of Multi-Purpose Portable Sieve Shaker" was carried out at the Department of Agricultural Process Engineering, College of Agricultural Engineering and Technology, Dr. Panjabrao Deshmukh Krishi Vidyapeeth, Akola. The study focused specifically on gram flour data from the thesis and on optimization of feed rate, sieve speed and sieve mesh size using Response Surface Methodology. The main objective was to obtain maximum flour yield with minimum flour loss.The study used the developed multi-purpose portable sieve shaker, which was intended as an economical and portable mechanization option for small-scale processing. The thesis identifies the three main objectives of the overall work as development of the machine, performance evaluation of the machine and cost analysis of the developed machine. The present paper is restricted to the gram-flour RSM component.

Materials and Methods

Experimental location and material

The research work was carried out in the Department of Agricultural Process Engineering, College of Agricultural Engineering and Technology, Dr. Panjabrao Deshmukh Krishi Vidyapeeth, Akola, Maharashtra. Gram flour used for the experiments was procured from the local market of Akola. The present manuscript uses the gram-flour experimental information reported in the thesis.

Plate 1: Multi-Purpose Portable Sieve Shaker Machine

Experimental design

Response Surface Methodology (RSM) was applied for optimization of the sieving process. A Box-Behnken Design for three independent variables at three levels was used. The design consisted of 17 trials, including five centre-point replications. The experimental data were analysed using Design-Expert software. The independent variables were feed rate (A), speed of sieve (B) and mesh size of sieve (C). The observed flour yield ranged from 95.30 to 97.70%. The highest observed yield was 97.70% at a feed rate of 2 kg/batch, sieve speed of 50 rpm and mesh size of 250 μm. Flour loss ranged from 2.30 to 4.70%, with the lowest observed loss of 2.30% at the same experimental combination. These observations indicate that the operating variables influenced the separation performance of the developed sieve shaker.

Effect of process and machine parameters on flour yield

The flour-yield response was fitted using a second-order polynomial model. The model F-value of 9.49 was significant (p = 0.0036), indicating that the fitted model explained a substantial part of the observed variation. Feed rate was the most influential linear term in the ANOVA, followed by sieve speed, while mesh size also contributed to the response. The quadratic terms A² and C² were significant, whereas the interaction terms were not significant at the 5% level. The non-significant lack of fit (F = 0.0815; p = 0.9667) indicated that the lack of fit was not significant relative to pure error. Regression equation in coded factors: Flour Yield = 97.04 -0.7750A + 0.2375B -0.1875C -0.0250AB + 0.0750AC -0.0500BC -0.3450A² + 0.0300B² -0.4200C²

Where, A = Feed rate (kg/batch) B = Speed of sieve (rpm) C = Mesh size of sieve (μm)

The thesis reports show that mesh size had a greater impact on recovery of flour as compared with feed rate and sieve speed.

Effect of process and machine parameters on flour loss

Flour loss ranged from 2.30 to 4.70% within the experimental combinations. The minimum observed flour loss was 2.30%. The fitted second-order response model showed the influence of feed rate, sieve speed and mesh size on flour loss. The thesis reports that mesh size had greater impact on flour loss compared with feed rate and sieve speed.

Design-Expert®

Numerical multi-response optimization

Numerical multi-response optimization was performed using Design-Expert software. The independent variables were kept within their experimental ranges, while flour yield was maximized and flour loss was minimized. The optimization criteria were feed rate 2-6 kg/batch, sieve speed 40-50 rpm, mesh size 200-300 μm, flour yield 95.3-97.7% and flour loss 2.3-4.7%. Each factor and response was assigned a desirability importance of 3 in the thesis. The selected numerical optimum reported in the thesis was feed rate = 2.000 kg/batch, sieve speed = 50.000 rpm and mesh size = 250 μm, with predicted flour yield = 97.419%, predicted flour loss = 2.581% and desirability = 1.000. This selected solution is retained here because it is the selected solution explicitly reported in the thesis.

Conclusion

The multi-purpose portable sieve shaker was evaluated for sieving of gram flour using Response Surface Methodology. Three process and machine variables-feed rate, sieve speed and mesh size-were optimized using a 17-run Box-Behnken Design. The observed flour yield ranged from 95.30 to 97.70%, while flour loss ranged from 2.30 to 4.70%. The flour-yield model was significant with R² = 0.9242 and adequate precision of 10.9045, and the lack of fit was non-significant. Numerical multi-response optimization selected a feed rate of 2.000 kg/batch, sieve speed of 50.000 rpm and mesh size of 250 μm. At these conditions, the predicted flour yield was 97.419% and predicted flour loss was 2.581%, with desirability of 1.000.