Cluster computing based optimization techniques for control of electrical drives

In recent times, the applications of clustering, parallel systems and distributed systems are widely used in many areas and there may exist an enormous amount of tasks which require parallel operations to improve the efficiency and performance of the system. One of the applications is control of flyback converter which is the simplest of the DC to DC converters with galvanic isolation between the input and the output sides provided by an isolation transformer. Flyback converters comprises of a power electronic switch, a set of diodes and a transformer in addition to a filter capacitor. The increased number of nonlinear elements along with a number of energy storage elements makes the design of an appropriate control system makes it challenging. This challenge is overcome by performing the functionalities simultaneously in different virtual machines (VM) considered as a single system. Considering the basic transfer function between the manipulated parameter which is the duty cycle and the controlled parameter which is the output voltage the forward and the reverse transfer function between these two parameters are estimated using MATLAB SIMULINK. The basic transfer function is identified and function is performed in a master VM whereas internal model controller (IMC) performs its functionalities in slave VM. Finally, the control system performance of the IMC is compared against the performance of a PI controller tuned with the particle swarm optimization based tuning procedure is analyzed using virtual machines.