Optimization of load distribution in washing machines using bio-inspired computational methods

This thesis deals with the optimization of the laundry load distribution within front loaded washing machines in order to avoid vibration noise and, in consequence, increase the performance and duration of such systems. The washing machine is programmed to follow several cycles in order to clean the laundry. The higher levels of vibration are reached when the motor executes the drying cycle where the laundry will spin at a very high speed in order to remove the water by using the generated centrifugal force. The aim of this thesis is to find a method which, when used, obtains a new distribution scheme that will evenly distribute the laundry load. This scheme will reduce the vibration with a low cost for the implementation since the change will be made only in the software of the system. In order to evaluate a possible solution scheme, a hardware-in-the-loop (HIL) set up was created. The HIL set up makes possible to avoid the complex modeling of the system as the washing machine itself will run each scheme and give a real time feedback of the unbalance. To improve the load distribution a novel approach is explored by using Evolutionary Algorithms that have been proved to solve problems with similar complexity involving a high number of variables. A scheme was obtained after the execution of the evolutionary algorithm that achieved a better performance of 12% less unbalance with 9.8% lower standard deviation than the current distribution scheme used commercially. With these results the validation of this method to obtain optimal distribution schemes is demonstrated.