The hazards of unexploded ordnance threaten the increasing marine construction activities nowadays, which increases the importance of unexploded ordnance detection. Research has shown that transient electromagnetic methods can successfully be used to detect unexploded ordnance on land. New equipment is being developed to make marine unexploded ordnance detection also possible. This study aims to determine which targets can be detected and which not in a marine environment through a numerical feasibility study. Building on an existing geophysical simulation framework, it asks: Under which conditions can we detect a conductor on or below the seafloor using a time domain loop source? Through the three-dimensional modelling of Maxwell’s equations, responses were computed for hollow rectangular targets of different burial depths, sizes, wall thicknesses, andwall conductivities. For the analysis of these responses two quantities were introduced, a net effect and a measurability. Evaluation of these quantities demonstrated the individual impact of the tested parameters on these quantities as well as the relative significance of the influence of these parameters. The results included derived relations for the influence of individual parameters on the net effect, as well as limits on the measurability of targets. Arectangular conductor of 0.1 by 0.1 by 0.4 metres or smaller with a wall thickness of 10 millimetre, buried more than 2 metres under the seafloor is not measurable under the noise assumptions made. The relative significance of the parameters was found to be from most to least significant: burial depth, size, wall conductivity, and wall thickness.

In open-pit mining on a bench level, the dig-limit optimization problem is deciding whether to classify a Selective Mining Unit(SMU) as waste or whether to classify it as ore. SMU’s with high ore grades are profitable for the mining operation, but due to equipment size a whole frame of SMU’s has to be classified as either ore or waste. In this thesis the dig-limit optimization problem will be solved using Mixed Integer Linear Programming (MILP). The model proposed in this thesis takes the frame constraint and the average grade constraint into account. The main results of the MILP model are in table 1. The results show that MILP can successfully be used to determine the optimal value of a mining operation, however the determination of the dig-limit for larger grids leads to exponential computation time.