Optimization of an air drying process for Artemisia absinthium leaves using response surface and artificial neural network models F. Karimi a , S. Rafiee b , A. Taheri-Garavand b, *, M. Karimi b a Department of Food Engineering, Faculty of Agricultural, University of Tabriz, Tabriz, Iran b Department of Agricultural Machinery Engineering, Faculty of Biosystems Engineering, University of Tehran, Karaj, Iran 1. Introduction Artemisia absinthium is commonly called wormwood and is locally known as ‘Afsentin’ in Iran. The plant can easily be cultivated in dry soil. They should be planted under bright exposure in fertile, mid-weight soil. It prefers soil rich in nitrogen. It can be propagated by growth (ripened cuttings taken in March or October in temperate climates) or by seeds in nursery beds. It is an ingredient in the liquor absinthe, and also used for flavouring in some other spirits and wines, including bitters, vermouth and pelinkovac. It is also used medically as a tonic, stomachic, antiseptic, antispasmodic, carminative, cholagogue, febrifuge and anthelmintic. In the Middle Ages it was used to spice mead. It is also an additional ingredient to mint tea in moroccan tea culture. The leaves and flowering tops are gathered when the plant is in full bloom, and dried naturally or with artificial heat. Its active substances include silica, two bitter elements (absinthine and anabsinthine), thujone, tannic and resinous substances, malic acid, and succinic acid. Its use has been claimed to remedy indigestion and gastric pain, it acts as an antiseptic, and as a febrifuge. For medicinal use, the herb is used to make a tea for helping pregnant women during pain of labor. A dried encapsulated form of the plant is used as an anthelmintic. A wine can also be made by macerating the herb. It is also available in powder form and as a tincture. The oil of the plant can be used as a cardiac stimulant to improve blood circulation. Pure wormwood oil is very poisonous, but with proper dosage poses little or no danger. Wormwood is mostly a stomach medicine [1–5]. Drying is defined as a process of moisture removal due to simultaneous heat and mass transfer. It is also a classical method of agricultural product preservation, which provides longer shelf- life, lighter weight for transportation and smaller space for storage [6]. Knowledge of heat and moisture transport is basis of process design, energy savings, and product quality. Determining moisture transport parameters for drying models are of particular interest for efficient mass transfer analysis and reproducibility of quality-controlled products. The continuously changing condi- tions along the period of the drying process make it difficult to determine the time duration of the process, and the most suitable values for the conditions to accomplish a successful drying process [7]. The most affecting factors related to the air drying are the air drying temperature, the air drying relative humidity and the air drying velocity in addition to the product initial moisture content [8]. Response surface methodology is a series of experimental design, analysis, and optimization techniques that originated in the work by Box and Wilson in 1951 [9]. The main idea of response surface methodology is to optimize an unknown and noisy function by means of simpler approximating functions that are valid over a small region using designed experiments. By moving Journal of the Taiwan Institute of Chemical Engineers 43 (2012) 29–39 A R T I C L E I N F O Article history: Received 6 December 2010 Received in revised form 11 March 2011 Accepted 4 April 2011 Available online 22 September 2011 Keywords: Artificial neural networks Optimization Response surface methodology Drying Artemisia absinthium A B S T R A C T Back-propagation artificial neural network and response surface methodology were used to investigate estimation capabilities of these two methodology and optimization acceptability of desirability functions methodology in an air drying process. The air temperature, air velocity and drying time were selected as independent factors in the process of drying Artemisia absinthium leaves. The dependent variables or responses were the moisture content, drying rate, energy efficiency and exergy efficiency. A rotatable central composite design as an adequate method was used to develop models for the responses in the response surface methodology. In addition to this isoresponse contour plots were helpful to predict the results by performing only limited set of experiments. The optimum operating conditions obtained from the artificial neural network models were moisture content 0.15 g/g, drying rate 0.35 g water/(g h), energy efficiency 0.73 and exergy efficiency 0.85, when air temperature, air velocity and drying time values were equal to 0.27 (47.3 8C), 0.02 (0.906 m/s) and 0.45 (10.35 h) in the coded units, respectively. ß 2011 Published by Elsevier B.V. on behalf of Taiwan Institute of Chemical Engineers. * Corresponding author. Tel.: +98 261 2801011; fax: +98 916 9795783. E-mail address: amin.taheri49@gmail.com (A. Taheri-Garavand). Contents lists available at ScienceDirect Journal of the Taiwan Institute of Chemical Engineers jou r nal h o mep age: w ww.els evier .co m/lo c ate/jtic e 1876-1070/$ – see front matter ß 2011 Published by Elsevier B.V. on behalf of Taiwan Institute of Chemical Engineers. doi:10.1016/j.jtice.2011.04.005