Mechanical Dewatering of Cassava Mash O. O. Ajibola ABSTRACT T O design appropriate presses for dewatering cassava mash, the dewatering characteristics of the mash need to be known. This paper reports on the factors that are important in the dewatering of cassava mash. The final moisture content of dewatering mash was found to be affected only by the applied pressure. The solid content of the expressed liquid did not change significantly during dewatering and was unaffected by such factors as the applied pressure, screen porosity and material depth. Mathematical models are presented which predict dewatering responses. INTRODUCTION Cassava (Manihot Utilissima) is consumed in West Africa mostly in the form of a fermented, semi- dextrinized meal called gari. Gari is prepared by peeling cassava tubers, grating them into mash, fermenting and dewatering the mash to a moisture content of about 100% dry basis (d.b.). The dewatering mash is then partially gelatinized, dried to a moisture content of about 12% d.b. and milled. The dewatering process is one of the most critical processes in gari production in that it influences greatly the gelatinization process and the cost of drying. In the traditional method of gari production, which is still the method used for producing most of the gari consumed in West Africa, dewatering is accomplished by putting the mash in hessian sacks and placing stones or some heavy material on the sack for a period of 48 to 90 h. During this time, fermentation and dewatering occur simultaneously. In industrial plants, dewatering is carried out after fermentation is complete with hydraulic or screw presses that accomplish the task in about 1 h. These presses have high initial and maintenance costs and are therefore not economical for small scale farmers. Since these farmers produce most of the gari, there is need for a low cost, simple and efficient machine for dewatering cassava. As such, the dewatering characteristics of cassava mash need to be identified. No such information could be found in the literature. The objectives of this study were to identify the factors that are important in the dewatering of cassava mash and to develop mathematical models to predict dewatering responses. DEWATERING MODELS Mechanical dewatering has been described as primarily a problem of fluid flow through porous media (Straub and Bruhn, 1978). Flow (q) through porous media of permeability (K) is described by Darcy's law as dx dp •[1] where — is the pressure gradient, dx Permeability depends on the porosity of the material. As the dewatering process continues, there is a deformation and creep of the material leading to a reduction in porosity and thus in the permeability. Deformation and creep during dewatering also reduces the material's thickness, so that at constant pressure, the pressure gradient increases with time. Bartlett et al. (1974) observed that those changes make it difficult to model dewatering using Darcy's law. Mechanical dewatering is similar to filtration which is the removal of solid particles from a fluid by passing the fluid through a filtering medium on which the solids are deposited. Perry and Chilton (1973), stated that filtration usually results in the formation of a layer (or cake) of solid particles on the surface of the filtering medium. Once this layer is formed, its surface acts as the filter medium, solids being deposited and adding to the thickness of the cake while allowing passage of the clear liquor. Thus, the overall resistance to flow of filtrate is equal to the sum of the cake resistance and the filtering medium resistance, with the filtering medium resistance being important only during the early stages of filtration. The cake is, therefore composed of a bulky mass of particles of irregular shape, permeated with small channels. The flow of filtrate through the channels is always streamline and may therefore be represented by Poisseuille's equation, which may be adapted in the following form: dV£ Adt AP /i[a(-) + r] A [2] Article has been reviewed and approved for publication by the Electrical and Electronic Systems Div. of ASAE. The study was performed with the grant URC No. 1425 HN from the University of Ife, Ile-Ife, Nigeria. The author is: O. O. AJIBOLA, Senior Lecturer, Agricultural Engineering Dept., University of Ife, Ile-Ife, Nigeria. Acknowledgments: I wish to acknowledge the contribution of Mr. Kola Egbedeyi at the data collection stage of the study. I wish to thank the International Institute of Tropical Agriculture (IITA) for the use of their computing facilities. where: Vf = T = A = M = P = volume of filtrate collected time area of filtering surface viscosity of filtrate total pressure drop across filtering medium and the cake deposited on it average specific cake resistance Vol. 30(2):March-April, 1987 © 1987 American Society of Agricultural Engineers 0001-2351/87/3002-0539$02.00 539