The electrification of the transport sector together with large renewable energy deployment requires powerful tools to efficiently use energy assets and infrastructure. In this framework, the forecast of electric vehicle demand and solar photovoltaic (PV) generation plays a fundamental role. This paper studies the impact of forecast accuracy on total electric cost of a simulated electric vehicles (EVs) charging station coupled with true solar PV and stationary battery energy storage. The optimal energy management system is based on the rolling horizon approach implemented in with a mixed integer linear program which takes as input the EV load forecast using long short-term memory (LSTM) neural network and persistence approaches and PV production forecast using a physical hybrid artificial neural network. The energy management system is firstly deployed and validated on an existing multi-good microgrid by achieving a discrepancy of state variables below 10% with respect to offline simulations. Then, eight weeks of simulations from each of the four seasons show that the accuracy of the forecast can increase operational costs by 10% equally distributed between the PV and EV forecasts. Finally, the accuracy of the combined PV and EV forecast matters more than single accuracies: LSTM outperforms persistence to predict the EV load (−30% root mean squared error), though when combined with PV forecast it has higher error (+15%) with corresponding higher operational costs (up to 5%).

Optimal behind-the-meter energy management often requires a day-ahead electric load forecast capable of learning non-linear and non-stationary patterns, due to the spatial disaggregation of loads and concept drift associated with time-varying physics and behavior. There are many promising machine learning techniques in the literature, but black box models lack explainability and therefore confidence in the models’ robustness can’t be achieved without thorough testing on data sets with varying and representative statistical properties. Therefore this work adopts and builds on some of the highest-performing load forecasting tools in the literature, which are Long Short-Term Memory recurrent networks, Empirical Mode Decomposition for feature engineering, and k-means clustering for outlier detection, and tests a combined methodology on seven different load data sets from six different load sectors. Forecast test set results are benchmarked against a seasonal naive model and SARIMA. The resultant skill scores range from −6.3% to 73%, indicating that the methodology adopted is often but not exclusively effective relative to the benchmarks.