An experimental and artificial neural network investigation on the laminar flow of magnetorheological fluids through circular pipes

Fluids can change their physical properties when they are exposed to magnetic fields. Magnetorheological (MR) fluids are classified as smart materials because their viscoelastic properties can increase by the application of the magnetic field. Accordingly, they are used in different engineering applications such as flow control and vibration damping. In this study, three different types of MR fluids flow in circular pipes with diameters of 10 and 15 mm and length of 300 mm were experimentally investigated with and without applying the magnetic field. An electromagnetic device was designed and manufactured in order to create a magnetic field induction for experiments. Throughout the experiments, the range of magnetic field induction value was B = 0-0.15 T, increased to 0.01 T. Based on the results obtained by the experimental study, it can be asserted that applying the magnetic field prompted an increase in the viscosity of MR fluids, leading to decreasing flow velocity. At B = 0.15 T, which is the highest value of the magnetic field, the flow velocity values dropped by 95%. Subsequently, the artificial neural networks algorithms are used in accordance with the obtained results to develop a correlation that clarifies the effect of the magnetic field on the flow velocity. The results show that the experimental and ANN models perform very similarly, and the ANN algorithm yields better results as a tool to predict the MR fluid flow behavior.