Droplet microfluidics is the field that allows the discretization of one fluid in another in channels of size less than a millimeter. Manipulation of these drops to effect different processes has been the focus of the microfluidics technology. Though the governing equations for flow inside a microchannel are linear, binary drop decisions at junctions and drop-drop interactions in a 2D microchannel make the flow problem non-linear and multi-scale. As a result drops form complex spatial and temporal patterns in a microchannel. To design these micro-devices it is important to understand how drops move and organize inside these channels. But the complex behavior of the drops inside the channel makes this a non-intuitive task. This calls for rational design approaches like the optimization based approach which demand simple models to represent flow of drops in microchannels. A CFD approach would be computationally intensive. Hence there is a need for a different modeling approach.
We propose an agent based approach to model drop motion inside microchannels. Each drop can be represented as an agent moving inside the microchannel interacting with other agents and the surroundings. These interactions are modeled using simple models or rules and incorporated inside a multi-agent framework that simulates the behavior of all the drops. We have applied the strategy to simulate self-organization of drops in a 2D microchannel. The interactions are modeled using ‘interacting drop traffic models’ and they are incorporated in a deterministic agent based framework. We show that the simulation strategy is able to predict the layering instability that results in patterns inside the 2D channel. We show how the pattern formation of drops can be related to the channel geometry. We also apply the agent based idea to study coalescence in a 2D drop assembly. One coalescence event can result in an avalanche of similar events resulting in the destabilization of the entire drop assembly. A stochastic multi agent approach is used where the propagation is modeled using a probabilistic rule. We are able to uncover the underlying auto-catalytic nature of the propagation and suggest design changes that would increase the stability of the assembly.
See more of this Group/Topical: Engineering Sciences and Fundamentals