ESR6 | Javier Lorente Macias

The research project focuses on the use of the adjoint method to optimize the performance of an inkjet printhead. This device is composed of many microchannels, each of which feeds a nozzle. An actuator placed on one of the walls of the microchannels is used to push ink through the nozzles, ejecting droplets of ink. However, this process also generates acoustic waves that propagate through the channels, increasing the acoustic energy of the system. This phenomenon affects the uniformity of the droplets. As a consequence, they may be expelled with different shapes and sizes, which compromises the quality of the product.

The purpose of this project is to find an optimal control law for the actuator that can damp the mechanical reverberations in the shortest time. However, this approach leads to an optimization problem with many design parameters, which is computationally expensive to solve with the traditional sensitivity analysis based on finite differences. Instead of that approach, we will use the adjoint method, which provides a cheap way to calculate gradient information in this kind of problems. This method will be applied to three-dimensional geometries that are being used in industry. We may also consider different optimization problems, for example shape optimization of the microchannels. In this case, the resulting geometries would passively damp the acoustic waves, without consumption of energy.

Finally, we will assimilate experimental data. The idea is to learn the model parameters and their uncertainties to obtain a more accurate representation of the physics of our problem.