Microwave vacuum drying of pomegranate peel: Evaluation of specific energy consumption and quality attributes by response surface methodology and artificial neural network

Abstract

Microwave vacuum drying of pomegranate peel was studied through this work and the drying process was modeled with the aid of response surface methodology (RSM) and artificial neural network (ANN) method. The drying experimental runs were performed on varied ranges of microwave power (175, 330, and 485 W) and vacuum pressure (10, 15, and 20 kPa) using a face-centered composite design. The influence of process parameters on five target responses, namely: total hydrolysable tannin (THT), color change (ΔE), maximum temperature evolved (T max ), effective moisture diffusivity (D eff ), and specific energy consumption (SEC) were analyzed. For microwave power and vacuum, the optimal conditions attained with aid of RSM and ANN were 260 W, 15.59 kPa, and 287 W, 19 kPa, respectively. The THT retention, ΔE, T max , D eff , SEC determined with RSM and ANN optimized values were 78.51 mg TAE/g of DW, 13.48, 67.02°C, 3.38 × 10 −6 m 2 /s, 20.02 kWh/kg and 90.32 ± 0.35 mg TAE/g of DW, 12.488, 64.83°C, 5.64 × 10 −6 m 2 /s, 20.01 ± 0.33 kWh/kg, respectively. The influence analysis explained that pomegranate peel dried at optimized ANN conditions of microwave power 287 W and vacuum 19 kPa possessed improved physicochemical properties and was efficient in terms of energy consumption compared to peel dried using RSM suggested conditions. Furthermore, the major criterion for evaluation between RSM and ANN tools in this study was the goodness of fit. The value for R 2 produced via RSM varied between 0.8354 and 0.9870, whereas subsequent R 2 in the case of ANN was considerably greater, varying between 0.9213 and 0.9956. The mean square error value displayed the minimum value in the case of ANN compared to the RSM model. These signified rather greater exactness and predictive capability of the ANN tool. Fourier transform infrared analysis established the existence of peaks between 3,308.92 and 908.89 cm −1 and high-performance liquid chromatography showed that pomegranate peel produced using microwave vacuum drying contains a good amount of gallic acid and tannic acid derivative compounds. Scanning electron microscopy elucidated that the combined microwave vacuum effect resulted in the formation of bigger pores on the peel surface ensuing in elevated extraction of THT at reduced energy consumption.

A comparative assessment of drying kinetics, energy consumption, mathematical modeling, and multivariate analysis of Indian borage (Plectranthus amboinicus) leaves

“…This was in accordance with the fact that higher drying rate was achieved with higher temperature, thereby reducing total drying time. Similar results have been observed by Elnjikkal Jerome and Dwivedi (2022) for drying of pomegranate peel, where higher temperature resulted in generation of higher energy due to more heat liberation, which enhanced the activity of water molecules in the sample, thereby resulting in quicker drying. Similarly for MD, the total drying time was 1080, 570, and 390 s (i.e., 18, 9.5, and 6.5 min) for microwave power of 200, 400, and 600 W, respectively.…”

“…D eff was determined by the method of slopes (Elnjikkal Jerome & Dwivedi, 2022), and the slope ( k ) was obtained from the graph plotted between ln(MR) versus drying time (Figure 1g–i). Higher D eff values were observed for MD (Table 1).…”

Modeling and optimization of osmo‐sonicated dehydration of garlic slices in a novel infrared dryer using artificial neural network and response surface methodology

“…Moreover, small variation was observed in statistical parameters in the RSM and ANN models, which indicates that both models have the capability for good prediction of the responses. These results are comparable with good prediction by ANN approach for microwave vacuum drying of pomegranate peel (Elnjikkal Jerome & Dwivedi, 2022).…”

“…Limited studies have been conducted on optimization and prediction through RSM and ANN models during the drying of ultrasonicated osmotic dehydration for aonla slices (Mehta et al, 2021), dragon fruit slices (Bhagya Raj & Dash, 2022), infrared drying of mulberry (Golpour et al, 2020). Moreover, RSM and ANN approaches were applied for optimization and modeling of the drying process on pomegranate peel drying by microwave vacuum (Elnjikkal Jerome & Dwivedi, 2022) and Aonla ( Phyllanthus emblica L .) slices (Mehta et al, 2021) ensured that the ANN modeling approach has a higher prediction capability.…”

“…Limited studies have been conducted on optimization and prediction through RSM and ANN models during the drying of ultrasonicated osmotic dehydration for aonla slices (Mehta et al, 2021), dragon fruit slices (Bhagya Raj & Dash, 2022), infrared drying of mulberry (Golpour et al, 2020). Moreover, RSM and ANN approaches were applied for optimization and modeling of the drying process on pomegranate peel drying by microwave vacuum (Elnjikkal Jerome & Dwivedi, 2022) and Aonla (Phyllanthus emblica L.) slices (Mehta et al, 2021) ensured that the ANN modeling approach has a higher prediction capability. However, fluidized bed drying of rice (Nanvakenari et al, 2021), fish drying by resistance heating based solar dryer (Richa et al, 2022), cocoyam slices (Onu et al, 2022), konjac drying (Zeng et al, 2020), apple drying in a convective and microwave dryer (Rasooli Sharabiani et al, 2021) reported that RSM demonstrated higher predictive capability than the ANN tool.…”