Droplet size prediction in a microfluidic flow focusing device using an adaptive network based fuzzy inference system
- PDF / 1,874,131 Bytes
- 12 Pages / 595.276 x 790.866 pts Page_size
- 115 Downloads / 248 Views
Droplet size prediction in a microfluidic flow focusing device using an adaptive network based fuzzy inference system Sina Mottaghi 1 & Mostafa Nazari 1 & S. Mahsa Fattahi 1 & Mohsen Nazari 1,2 & Saeed Babamohammadi 1
# Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Microfluidics has wide applications in different technologies such as biomedical engineering, chemistry engineering, and medicine. Generating droplets with desired size for special applications needs costly and time-consuming iterations due to the nonlinear behavior of multiphase flow in a microfluidic device and the effect of several parameters on it. Hence, designing a flexible way to predict the droplet size is necessary. In this paper, we use the Adaptive Neural Fuzzy Inference System (ANFIS), by mixing the artificial neural network (ANN) and fuzzy inference system (FIS), to study the parameters which have effects on droplet size. The four main dimensionless parameters, i.e. the Capillary number, the Reynolds number, the flow ratio and the viscosity ratio are regarded as the inputs and the droplet diameter as the output of the ANFIS. Using dimensionless groups cause to extract more comprehensive results and avoiding more experimental tests. With the ANFIS, droplet sizes could be predicted with the coefficient of determination of 0.92. Keywords Microfluidics . ANFIS . Droplet generation . Fuzzy based neural network
1 Introduction Currently, microfluidic and micro-droplets have wide applications in various areas such as biomedical engineering, chemical engineering, and medicine (Lashkaripour et al. 2018; Jung and Oh 2014; Song et al. 2003). Micro-droplets provide a large advantage for not only the engineers but also for the pharmacists and therapists because of its high precision, accuracy, sensitivity and fast reaction time and its small size (Ray et al. 2017). Droplets have a confined space which can be used as an ideal reactor for biochemical and chemical processes (Nguyen et al. 2010). However, the field of microfluidics has not been deployed in the life sciences due to the costly and time-consuming process of manufacturing and the complex and nonlinear dynamics of the two phase flow (i.e. two immiscible fluids such as oil and water inside the microfluidic channel). Hence, presenting models which can investigate the role of different effective parameters on the droplet size is very * Mostafa Nazari [email protected] 1
Faculty of Mechanical and Mechatronics Engineering, Shahrood University of Technology, Shahrood, Iran
2
Visualization and Tracking Laboratory, Shahrood University of Technology, Shahrood, Iran
attractive. These models can save time and money by avoiding experimental tests. The first concept that exists here is the way of generating droplets. A large variety of methods with some advantages and drawbacks has been presented so far for this purpose (Park et al. 2011; Mastiani et al. 2019; Murshed et al. 2009). It can be said that electricity has had the most use in this field (Chong et al
Data Loading...
