Distributed energy resource management considering limited distribution network monitoring

The evidence of climate change and the associated agreements on carbon emission reduction targets are motivating an increasing number of distributed energy resources (DER) to be installed in the distribution network. This is likely to cause violations of system constraints in the distribution network. Furthermore, limited monitoring in the distribution network leads to uncertainty, in turn increasing this risk. However, the flexibility introduced by DER could support the distribution network. DER aggregators (DERAs) have been proposed to manage multiple small-scale DER. Significant research has been conducted on the impact of DER and the methods to manage them, but less has been published regarding uncertainty handling in real time considering the limited data available to a DERA in a realistic network. To address these limitations, methods are proposed that can be used to achieve DER management by a third-party DERA considering the uncertainty in the distribution network. The impact of DER will be evaluated in a shared parking facility distribution network using the University College Dublin distribution network as a test network. As electric vehicles gain popularity in Ireland, it is a good example of DER that, if left unmanaged, could cause significant concerns in the distribution network. Motivated by the fact that the distribution network will likely become congested with the increase in DER, the flexibility available due to DER in a distribution network is considered. The proposed method considers uncertainty while including all available data. A probabilistic model is developed that includes historical data and any available measurements in order to define a feasible region in which any changes to the DER output respect system operating constraints. This creates a more precise model that more accurately matches the state of the network and reduces the uncertainty. In order to manage DER in real time, it is imperative that a DERA know the impact of changing DER output prior to dispatch. A method is proposed that estimates the voltages at the points of connection of DER based on a desired change in DER active power generation in real time. This allows DERAs to mitigate the risk of violating system constraints while managing DER. This method relies only on information reasonably available to a DERA. No knowledge of the network is required. The method uses a fully connected feedforward neural network to estimate the voltages at the points of DER connections after desired changes in each DER active power generation. The proposed method is validated for both single-phase and three-phase voltage estimations and shows superior performance compared to a voltage sensitivity approach.