Due to its impact on household energy use and the adoption of renewable energies, the intelligent management of the power consumption of electric vehicles (EVs) is of great relevance. In the context of widespread clean energy adoption and growing environmental concerns, generating incentives through discounted rates for intelligent residential EV power consumption requires algorithms capable of measuring loads in a disaggregated manner. The deployment of smart meter networks offers the possibility of applying machine learning techniques to estimate EV residential consumption. This work presents an efficient algorithm for the Non Intrusive Load Monitoring (NILM) of EV consumption, which is an adaptation of a method previously proposed for high-powered water heaters. Its performance is compared with methods based on deep neural networks. Results from an actual power demand dataset are discussed, and a comparative analysis is carried out against billing rules based on time slots and historical power consumption data.