Minimization of Voltage Deviations, Power Losses and Control Actions in a Transmission Power System

A multi-objective genetic algorithm, based on NSGA- II, is implemented to find an optimal condition of minimum voltage deviations, minimum power losses and minimum number of control actions of a transmission network system. The system used as model is an IEEE 30-bus system. Generators and transformers with off-nominal tap ratio are the devices to be controlled. Different probabilities of mutation factors are compared and it is proved that a more important mutation factor can improve the velocity of convergence without getting into a random search. The algorithm is compared with other published results and a successful performance of the implemented algorithm is proved. Strategies for the operation of power transmission networks have been increasingly influenced by a philosophy of minimization of economic losses and maximization of the quality of the delivered energy. This is a policy for optimal operation that has become a common goal for electric power companies. In the near future, operating the system at loading limits will be a normal practice for power engineers. But, for security and economic reasons it is very important to develop ever-better control algorithms. Control of voltage levels is one of those algorithms that have been developed and need to keep developing in order to guarantee a secure and profitable exploitation of the network. In essence, the control of voltage levels is accomplished by controlling the production, absorption, and flow of reactive power at all levels in the system. The devices used for this purpose are series and shunt capacitors, shunt reactors, synchronous generators, tap- changing transformers and others. The voltage control algorithm should guarantee that voltage levels at all buses are within acceptable limits and also bring a reduction of I 2 R and I 2 X losses (1). Many algorithms have been proposed for the control of voltage levels. Initial proposals were based on classic methods of optimization but the complexity and diversity of objectives and constraints make very difficult to find global optimal solutions. More recent approaches based on evolutionary strategies (ES) have opened new ways to solve the problem with promising results. Among the ES, Genetic Algorithm