The optimization of network capacity utilization is becoming increasingly important in an elec- trical distribution network with the variation in load demand due to seasonal changes and uptake of new technologies such as photovoltaics, batteries, electric vehicles and distributed generation. The proliferation in the use of these technologies introduces uncertainties in future load demands and increases the risk of asset stranding. This then represents an opportunity for network plan- ners and operators to re-evaluate their existing infrastructure with the aim to optimize the electrical components operation and return on investment.

This thesis introduces a novel method to evaluate and optimise the network capacity utilization for network planning, operation and management. A predictive asset rating model for 33kV electrical distribution has been developed and implemented in Unison Network Limited as part of this research. The thermal behaviour of the electrical components to its surrounding was studied utilising sensors installed along the circuits and based on historically captured data available from the public and at localised area. The predictive asset rating model performs network evaluation based on probabilistic method to evaluate the dynamic utilisation of the network and reliability in the event of changes to the operating conditions and the network configuration. The model was developed using MATLAB to carry out different simulation scenarios for network planning and operation.

The thesis begins by introducing the current state of electrical distribution networks. The rating and modelling of electrical components is then introduced, covering international standards and similar research in the area of study. The impact of environmental conditions and load demand to the available capacity of the network are then discussed. Results in the area of capacity planning, constraint analysis and network expansion planning is presented, highlighting the model’s ability to predict capacity utilisation and constraints on the network.