Clusters of galaxies are the largest objects to collapse under their own gravity and virialise. Cluster properties provide a unique way of constraining cosmological parameters. This is achieved using the mass function via the scaling relations. While the mass function can be directly related to the cosmology of the Universe, the scaling relations relate properties of clusters to their mass. The scaling relations are thus used to place clusters on the mass function by relating mass to another observational property.The simplest model of predicting the scaling relations is the self-similar model. Observational data show, however, that the self-similar model is insufficient, requiring additional processes which are not gravitational in origin. In this thesis, a new simulation has been studied to investigate the effects of these non-gravitational processes (specifically, feedback from active galactic nuclei) on the cluster scaling relations. The results from the scaling relations at redshift zero show a similar result to the observational data, in that the self-similar relations are broken. When examining the scaling relations at higher redshifts, the luminosity relations show an even larger departure from the self-similar predictions in both normalisation and slope. However, temperature and Y parameter relations show a slow tendency to the self-similar result in terms of the slope but not in the normalisation. Also found is a slight departure from self-similarity in the Mgas -M relation slope but not in the normalisation. I conclude that scaling relations can not all be self-similar, so a new model for the scaling relations is required.