Fluid systems

Fluid systems

Droplet detection and classification by Anshul Verma

Droplet detection and classification is a necessity for one to implement different droplet manipulation in micro/millifluidic network. To meet this need I have designed a sensor using a LDR (Light Dependent Resistor). The channel is sandwiched between the LDR and the light source. Thus when the drop passes under the LDR there is a change in the type of fluid present in between the LDR and the light source. This change of medium from continuous medium to droplets medium leads to change in intensity of light falling on the LDR. This intensity change leads to resistance change of the LDR. This resistance change of LDR is what leads to the signal generated by the sensor. The peaks in the sensor signal signifies drop.


The change in medium between the LDR and the light source leads to change in intensity and wavelength of light falling on the LDR. The change in intensity of the light is due to the change in absorption, reflection and transmission coefficients of the two fluids used to form drops. This change in intensity and wavelength is captured by the LDR in form of its resistance change. To keep track of this resistance change of LDR, I use LDR in a wheat-stone bridge with LDR being one of the branches of the bridge. Thus this change in LDR’s resistance is converted to the voltage change across the bridge. Which is then measured by using the Arduino. The Arduino’s output is then stored using MATLAB and ARDUINO software. The data which is obtained from the bridge is not amplified currently but can be amplified if required. This sensor’s data is very noisy and to make interpretations of this data I filter the data to remove the high frequency noise (mostly 50 Hz). I generally use weighted moving average filter. But any other form of high pass filter can also be used.

Arduino’s analog inputs are used to measure the DC voltage of the two nodes of Vo which are later differenced and stored using MATLAB. This differenced voltage is the signal corresponding to the sensors which has peaks which signifies presence of a drops under the sensor.

The feature of the peaks must be related to the features of the drops. I try to exploit this fact and use these features as the variables, trying to define different characters of drops like type (for classification), radius and velocity. Some of these features like peak width should ideally be independent of the characteristics like class of drop but we feed in all the possible features to the classifier because the classifier should able to distinguish between the relevant and non-relevant features. Thus the presence of non-relevant features should not be an issue.

 Droplet millifluidics experiments by Arun Sankar E M

Droplet millifluidics deals with the generation and manipulation of drops in channels having size of the order of a millimeter. The ease with which the device is fabricated, and the flow is visualized makes this a very convenient and attractive platform to carry out experiments. We fabricate PDMS millifluidic channel using metallic molds and capture the droplet motion using a normal camera. One of the attractive features of microfluidic/ millifluidic droplet generation is that it can produce monodisperse droplets. This ability to produce monodisperse droplets is exploited in areas such as reaction engineering where droplets are considered as microreactors, particle synthesis where droplet is polymerized by photopolymerization etc[1]. Now, these drops are allowed to interact in a 2D space. They interact and form multi layered patterns in a 2D diverging-converging millichannel. This is similar to that observed by Jose and Cubaud [2] in microchannel in similar geometry. The similarity in the droplet generation and droplet interactions in the two scales motivate to study if millichannels can be used as test stations before directly carrying out experiments in the rather expensive microchannels.


  • W. Engl, M. Tachibana, P. Panizza, and R. Backov, “Millifluidic as a versatile reactor to tune size and aspect ratio of large polymerized objects,” Int. J. Multiph. Flow, vol. 33, no. 8, pp. 897–903, 2007.
  • B. M. Jose and T. Cubaud, “Droplet arrangement and coalescence in diverging / converging microchannels,” Microfluid Nanofluid, no. 12, pp. 687–696, 2012.

 Droplet microfluidics- a systems view by Danny Raj M

Droplet microfluidics is the field of study which consists of producing and manipulating droplets inside very small channels for different applications that range from particle synthesis to DNA sequencing to soft computing. An experimentalist who fabricates a microchannel and observes the motion of droplets in the channel typically solves the forward problem which involves understanding the behaviour of droplets for a given geometry and operating condition. However, a design problem is an inverse problem as it involves identifying the geometry and operating strategy required to yield the desired behaviour of droplets in a microchannel. And the non-linear nature of droplet flow which results in collective droplet motion renders design non-intuitive. Hence, there is a need for a systematic approach to design which would involve posing the design problem as an optimization problem. However, the multi-scale nature of the flow problem makes the conventional computational fluid dynamics approach infeasible for design. And, without a computationally simple modelling strategy for droplet flow, it will be impossible to uncover the full potential of the droplet microfluidics technology. Hence, during my doctoral work, I proposed to address this problem by introducing the agent-based approach to simulate complex droplet flow and behaviour in microchannels.

Self organization of drops

Drops form ordered arrangements inside 2D microchannels. The complex traffic inside the channel is a result of the hydrodynamic interactions of the drops. We propose an agent based approach to model the system. The framework involves identifying the drop-level interactions using simple phenomenological models and a multi-agent simulation which pieces the entity-level information to predict the collective dynamics of the system.

List of Publications

  • M Danny Raj and R. Rengaswamy, “Understanding drop-pattern formation in 2-D microchannels: a multi-agent approach”, Microfluid. Nanofluidics, vol. 17, no. 3, pp. 527–537, Jan. 2014.
  • D. Raj and R. Rengaswamy, “Investigating Arrangement of Composite Drops in Two-Dimensional Microchannels Using Multi-Agent Simulations: A Design Perspective”, Ind. Eng. Chem. Res., vol. 54, no. 43, pp. 10835-10842, 2015.

Coalescence avalanches

An isolated coalescence event, in a concentrated emulsion flowing through a 2D microchannel, can trigger an avalanche of similar events which results in spontaneous destabilization of the droplet assembly. We propose a stochastic model to simulate the phenomenon

List of Publications

  • Danny Raj and R. Rengaswamy, “Coalescence of drops in a 2D microchannel: critical transitions to autocatalytic behaviour”, Soft Matter, vol. 12, no. 1, pp. 115-122, 2016
  • Danny Raj and R. Rengaswamy, “Averaged model for probabilistic coalescence avalanches in two-dimensional emulsions: Insights into uncertainty propagation”, Physical Review E, vol. 95, no. 3, 032608, 2017.


 Spinning Disk Atomization (SDA) by Sreejith C

Micro-encapsulation is the process of forming fine particles or microspheres with a core engulfed in a matrix. Spinning Disk Atomization (SDA) is a technique used for Micro-encapsulation. In SDA, liquid atomization is achieved by introducing the liquid to the surface of a high speed rotating disk. Micro-encapsulation finds applications in various fields such as pharmaceuticals, detergents and food. These ever growing fields demand a continuously operated set up to produce such encapsulated microspheres. In view of these, this research work is intended to develop a compact equipment set up for the continuous production of Alginate hydrogel microspheres by SDA at a high production rate.
This is achieved by reacting atomized Sodium Alginate aqueous solution with Copper Sulphate in a specially designed spinning disk equipment to form Alginate hydrogel microspheres. The slurry formed is treated in a downstream unit to separate the particles and the liquid is recycled to the reaction system to improve efficiency of operation.
The design is aimed to provide high efficiency operation with maximum flexibility in a compact manner. Once a stable operation of the units is established with the current encapsulation method, advanced encapsulation methods can be tried by making suitable alterations to the equipment set up.