Plug-in Electric Vehicle (PEV) user charging behavior has a significant influence on a distribution network and its reliability. Generally, monitoring energy consumption has become one of the most important factors in green and micro grids; therefore, predicting the charging demand of PEVs (the energy consumed during the charging session) could help to efficiently manage the electric grid. Consequently, three machine learning methods are applied in this research to predict the charging demand for the PEV user after a charging session starts. This approach is validated using a dataset consisting of seven years of charging events collected from public charging stations in the state of Nebraska, USA. The results show that the regression method, XGBoost, slightly outperforms the other methods in predicting the charging demand, with an RMSE equal to 6.68 kWh and R² equal to 51.9%. The relative importance of input variables is also discussed, showing that the user’s historical average demand has the most predictive value. Accurate prediction of session charging demand, as opposed to the daily or hourly demand of multiple users, has many possible applications for utility companies and charging networks, including scheduling, grid stability, and smart grid integration.

Plug-in Electric Vehicle (PEV) charging stations can have a significant influence on the electric grid and its reliability. In order to evaluate the influence of rapid penetration of PEVs and their charging infrastructure on the electric grid, and assess their potential benefits to the future micro-grid, statistical analysis of PEV charging behavior is essential for planning as well as running public charging stations in different areas. In this paper, 27,481 charging sessions from public charging stations in the state of Nebraska are analyzed to determine the relationship between energy consumption, time, and location of charging. A multinomial logistic regression model is presented to aid utility companies in anticipating the charging demand of PEV users, as well as the impact on the electric grid at increasing PEV penetration rates.

The power consumption of buildings over the course of each minute, hour, day and season plays a major role in how this load influences the Electric Power System voltage and frequency, and vice versa. This consumption is based on the building’s load component types, efficiencies, and how they consume power and react to changes in real time. Due to this complexity, standard full-building load models are typically voltage-invariant. This paper proposes a novel framework to transform these voltage-invariant building load models into fully time- and voltage-dependent load profiles using available data on the voltage sensitivity of individual load components. While a voltage-dependent building model could theoretically be generated from static load models of every component in a building, this approach faces two challenges: first, load models representing all load components are impractical to develop for all possible load component types; second, building energy consumption is never measured or modeled at the individual component level. The proposed framework compiles available component data in the form of static ZIP load model parameters, and maps them into the end use categories utilized by standard building modeling programs. The voltage sensitivity of each end use category is then bounded by the extrema of the component models within it. This framework is applied to a load profile case study representing the aggregate U.S. residential building stock. In addition to the minimum/maximum conditions, a load profile based on typical load composition and weighted ZIP parameters is generated for the same building stock. The results show that for a 10% drop in voltage, using the least sensitive ZIP parameters, active power is expected to be 3% to 14% lower than nominal, depending on the season and time of day. Using the most sensitive ZIP parameters, the active power is expected to be 9% to 20% lower than nominal, also depending on the season and time of day.

All-Electric Vehicles (EV s ), BATTERY-powered EVs (BEVs), and plug-in hybrid EVs (PHEVS) are gaining market share and increasing in popularity with the buying public because the battery range (longer) and cost (lower) have reached sweet spots, the charging infrastructure is more robust, and concern with global climate change is high. In 2013, only 100,000 EVs were sold in the United States, but by 2022, approximately 800,000 have been purchased. A similar growth is seen in EV supply equipment (EVSE), i.e., EV charging stations, with 19,742 documented EV charging station locations in the United States in 2013 to 50,054 documented EV charging station locations, with approximately 130,000 ports, by the end of 2022.

While the increase in EV use is a positive step towards embracing green technology, the heightened energy demands resulting from this rapid growth present major challenges to local energy grid load management. For context, the new EVs being deployed store approximately 100 kWh, about four times the daily electricity use of the average household in the U.S. Current local distribution grids do not have the capacity to accommodate these massively increased loads. For this reason, it is important that local utilities have a full understanding of the charging demand within a given grid. The main objective of this research is to deepen the understanding of charging behavior at the household level using real data. Specifically, data from existing 417 residential Level-2 charging stations, located in Omaha, Nebraska, USA, are collected and analyzed. The results show a clear pattern in user behavior for the starting time of sessions, as well as the connection duration, charging duration, and subsequent energy demand.