The generation volatility of photovoltaics (PVs) has created several control and operation challenges for grid operators. For a secure and reliable day or hour-ahead electricity dispatch, the grid operators need the visibility of their synchronous and asynchronous generators’ capacity. It helps them to manage the spinning reserve, inertia and frequency response during any contingency events. This study attempts to provide a machine learning-based PV power generation forecasting for both the short and long-term. The study has chosen Alice Springs, one of the geographically solar energy-rich areas in Australia, and considered various environmental parameters. Different machine learning algorithms, including Linear Regression, Polynomial Regression, Decision Tree Regression, Support Vector Regression, Random Forest Regression, Long Short-Term Memory, and Multilayer Perceptron Regression, are considered in the study. Various comparative performance analysis is conducted for both normal and uncertain cases and found that Random Forest Regression performed better for our dataset. The impact of data normalization on forecasting performance is also analyzed using multiple performance metrics. The study may help the grid operators to choose an appropriate PV power forecasting algorithm and plan the time-ahead generation volatility.

Peak-load management is an important process that allows energy providers to reshape load profiles, increase energy efficiency, and reduce overall operational costs and carbon emissions. This paper presents an improved decision-tree-based algorithm to reduce the peak load in residential distribution networks by coordinated control of electric vehicles (EVs), photovoltaic (PV) units, and battery energy-storage systems (BESSs). The peak-load reduction is achieved by reading the domestic load in real time through a smart meter and taking appropriate coordinated action by a controller using the proposed algorithm. The proposed control algorithm was tested on a real distribution network using real load patterns and load dynamics, and validated in a laboratory experiment. Two types of EVs with fast and flexible charging capability, a PV unit, and BESSs were used to test the performance of the proposed control algorithm, which is compared with that of an artificial-neural-network technique. The results show that using the proposed method, the peak demand on the distribution grid can be reduced significantly, thereby greatly improving the load factor.

The increased integration of light-duty electric vehicles (EVs) into low-voltage (LV) residential networks imposes capacity issues for the grid operators. For example, the uncoordinated and clustered charging of residential EVs can often overload grid assets, jeopardize network reliability, and violate local voltage constraints. This article proposes a coordinated management system for EVs in LV residential networks with power grid support functionalities to address grid overloading and local voltage constraints violation. The charging and discharging of EV batteries in the network are coordinated via a local EV aggregator. The coordination is realized using multiagent system architecture that provides the EV owners with full decision-making authority and preserves their privacy. The EV coordination and vehicle-to-grid (V2G) resource optimization of the EV aggregator is formulated as a mixed-integer programming-based optimization model to minimize the electricity cost for the EV owners based on a real-time tariff while complying with local grid constraints. The proposed methodology is evaluated via simulation of an LV residential network in Sydney, Australia with actual load demand data. The simulation results indicate the efficacy of the proposed strategy for the electricity cost reduction of the EV owners while mitigating grid overloading and maintaining desired bus voltages.