Distribution networks planning using decomposition optimisation technique

This study proposes an efficient planning method for distribution networks in order to minimise power loss and maximise reliability taking into account the distributed generation, transformers load tap changer and capacitors location and size. This multi-criteria model is subjected to several technical constraints and full AC-optimal power flow (OPF). The planning process is running by means of a multi-criteria analysis in which three objectives are considered: the minimisation of the system´s power loss; the minimisation of capacitor placement and size cost; and the minimisation of energy not supplied cost, maximising reliability. For solving this problem, decomposition techniques were chosen. This way, the problem was divided into two parts - one called the master sub-problem and another called slave sub-problem. The master sub-problem (decision sub-problem) determines the radial topology of the distribution network and was formulated as a mixed-integer quadratic constraint problem. The slave sub-problem is used to define the feasibility of the decision sub-problem solution by solving an OPF, giving information required to formulate the linear cuts and formulated as a non-linear programming problem. These cuts link the master and the slave sub-problems. The methodology is programmed in the software General Algebraic Modelling System and it is applied to a 70-bus distribution network.