Method of inferring operation status of distributed generation systems in distribution section

Recently, many distributed generation systems have been installed in power distribution networks. In the case of an extensive penetration of distributed generation systems, serious problems such as overvoltage at the customer receiving end as a result of surplus photovoltaic generation and islanding phenomena at a fault in a distribution line may occur. In particular, the distribution line may become overloaded during fault recovery because all distributed generation systems must stop operating instantly to prevent islanding at the fault; then, generation cannot be resumed instantly in accordance with the grid interconnection code in Japan. However, distribution network operators can prevent overload if the operation status of distributed generators can be estimated in advance. As a first step of inferring the operation status of distributed generators, we developed two methods of inferring the operation status of distributed generators at the customer receiving end. One was a method for low-voltage customers, in which the change in power factor measured at the customer receiving end was used. The other was a method for medium-voltage customers, in which both the active power and power factor measured at the customer receiving end were used. These two methods could infer that the distributed generators are “in operation” or “not in operation” at the customer receiving ends. However, the operation status in a distribution section, which is sectioned by two pole-mounted gas switches, should be inferred from the viewpoint of distribution operation. As the active power and power factor in the distribution section are affected by the loads of other customers, it is difficult to adopt the method for medium-voltage customer´s receiving ends using the active power and power factor. This method could not infer the numbers of operating distributed generators. Therefore, we develop a new method of inferring both the status and number of distrib- - uted generators. This method uses the change in phase current unbalance due to the distributed generator operation. We apply this method to three distribution lines including medium-voltage customers. The correct answer rates of inference of the status and number of operating distributed generators for three different distribution sections were ranged from 95.7% to 98.1%, and from 86.1% to 97.0%, respectively..