Detonation can be described as a phenomenon resulting in formation of a self-sustaining detonation wave (shock wave accompanied with a chemical reaction zone) moving with supersonic speed (up to 10 km/s) and extremely high pressure and temperature (up to 40 GPa and 6000 K) in nanoseconds. There are two generally accepted detonation theories based on conservation laws and hydrodynamic flow models: Chapman-Jouguet (CJ) theory, which assumes instantaneous chemical reactions with no chemical reaction zone, and Zeldovich-von Neumann-Doering (ZND) theory, which assumes existence of a chemical reaction zone of definite length and duration time. Improvements in experimental techniques and understanding of energy transfer and kinetics in chemical reaction zone have directly impacted improvements in detonation models and understanding of nonideality. Nonideal explosives are considered those whose behaviour can not be accurately described with a simple CJ theory. Experimentally obtained detonation velocity and pressure of such explosives are considerably lower than values calculated by CJ theory. On top of that, detonation velocity shows a very high dependence on charge radius and existence of confinement, which is not the case for ideal explosives. Typical example of nonideal explosives is ANFO, one of most used commercial explosives. Main cause of nonideality is relatively long duration of chemical reactions in a chemical reaction zone (in microseconds compared to nanosecond scale observed in ideal explosives) directly resulting in a wide chemical reaction zone (in tens of millimetres). Besides that, these slow chemical reactions also result in other nonideal characteristics such as curves detonation wave front, nonlinear dependence of detonation velocity on initial explosive density, dependence of detonation parameters on charge radius and existence of confinement and partly reacted explosive at the end of a chemical reaction zone. To satisfactory describe nonideal detonation, nonideal detonation model, used for numerical modelling of nonideal explosive behaviour, has to be based on a nonideal detonation theory, as well as take into account reaction rate law (or chemical kinetics), radial expansion of detonation products and equations of state of unreacted explosive and detonation products. Many scientists have devoted their time to researching nonideal detonation models which would be able to satisfactory and accurately describe the behaviour of highly nonideal explosives such as ANFO. However, there is still no generally accepted nonideal detonation model as there is still up to 50% discrepancy between experimentally measured and theoretically calculated detonation velocity. Other major problems in dealing with numerical modelling of nonideal detonation are reliance on empirical data, use of incomplete equations of state, too simple or too complicated reaction rate laws and complete absence of theoretically based confinement model. This thesis deals with improvements of nonideal detonation model based on one of the most used nonideal detonation theory, Wood-Kirkwood slightly divergent flow theory, incorporated in EXPLO5 thermochemical code. Main goals of this thesis are: • Experimental determination of detonation velocity as a function of explosive charge radius to obtain an analytical link between two parameters, increase the reliability of the model and decrease the future need for empirical input data. • Development of a reliable and theoretically based temperature and/or pressure-based reaction rate law which can adequately describe ANFO behaviour. • Experimental determination of curvature radius and development of radial expansion model calibrated with experimental curvature radius data. Scientific impact of this thesis, other than improvements made to the reaction rate and radial expansion model in nonideal detonation model which make it more predictable, is the contribution to better understanding of chemical reaction zone structure and processes for heterogeneous ANFO explosive.