FE-modeling and validation of a washing machine : Using experimental modal analysis and spin up test

Abstract

In this master thesis the dynamics in the steel shell of a washing machine from ASKO Appliances AB has been studied. The thesis is the third in a series where the first studied the behavior of a side plate and the second studied the dynamics in half of the steel shell parts. In this work all parts are first studied separate and then assembled to the steel shell. Additionally a model will be made of the parts needed to simulate the spin up of a washing machine. The goal of the thesis work is to make a model of the steel shell that correlates well with the real steel shell and to be able to simulate the forces and accelerations in the shell that arises from a spin up in the drum of the washing machine. Every detail has been modeled using finite elements and validated using experimental modal analysis. The parts already studied will also be modeled in order to match the results from the previous year's thesis works. The results show that each separate part correlates well in most modes when comparing mode shapes, eigenfrequencies and frequency response functions. The models are assembled into the steel shell and validated in the same manner as each separate part. The result from these studies shows that a good correlation is obtained in the steel shell up to about 60 Hz. Above that, the mode shapes are getting more complex and harder to correlate. There are some differences in eigenfrequencies between the model and the real steel shell. The problem with the side plate correlation of previous thesis works have been found to have minor importance when the steel shell had been fully assembled. The model used to calculate accelerations and forces in the machine during rotation of an unbalanced loaded drum showed a rather poor correlation. The reason is that higher order harmonics have more impact on the vibrations in the shell than first thought. The forces under the feet and the accelerations in the side plate were the results that matched the measured data best.