Many process industries deal with particulate materials, often in the form of bulk mixtures with considerable variations in particle size and shape. Degradation (breakage) of particles during transport, handling and processing presents a serious handling problem – improved understanding of degradation processes is a highly desirable aim. When bulk assemblies are dense-phase in nature, degradation of particles is likely to be significantly influenced by the presence of surrounding particles. Quantifying such influence allows better modeling and predictive capability. Techniques involving matrix equations are potentially useful modelling tools, already successfully used in predicting degradation of (lean-phase) assemblies, where the influence of particles on each other is probably negligible. The motivation of this research is to pioneer use of breakage matrix techniques as more versatile tools to predict degradation in dense-phase assemblies (significant interacting mixture breakage conditions). The principal issue with the breakage matrix approach is the amount of necessary data and the difficulty in obtaining that data, particularly when there is intra-mixture influence as described above. A technique is pioneered, based on experimental data obtained from compression and impact tests of mixtures of assemblies, by which breakage matrices can be calculated from only the input and output particle size distributions for a degradation process. The results reveal that the degree of intra-mixture influence can be quantified by recourse to fairly straightforward analytical expressions of coefficient of interaction, and indicate ways in which breakage matrices can be inferred from scarce available data. The results show that the breakage matrix approach to be a more promising tool for modeling and predicting interacting and non-interacting mixture degradation than has been appreciated to date. The significance of the semi-empirical correlations established for interacting and noninteracting mixtures could further be demonstrated in industrial applications such as through systematic sampling in pneumatic conveyors and in silo filling and discharge. Industrial field data could further be compared with the results of controlled experiments in model compaction, attrition and shear cells demonstrated in the thesis to pave the way for an universal approach.