Non-Intrusive Load Monitoring (NILM) is the process of deducing/identifying appliances and their energy consumption in a household. With NILM, we enable energy disaggregation – Decomposing power consumption levels at the appliance level from the aggregate consumption at the house level. Since NILM is one of the most promising techniques for analyzing and upgrading our electricity grid in a smart and scalable manner, we at Plexflo built an Open-Source library datastream that aids researchers and engineers to try our Deep Learning models for the detection of EVs (Electric Vehicles) charging events from smart home meter data.
We are thrilled to announce the release of Plexflo’s python library : datastream
We have the first of many sub-modules called the datastream ready to be used by the community. datastreamwill enable thousands of researchers, engineers, and energy companies to build intelligent insights, prototype, and test our latest technology before wider scale adoption! And for this release, we are focusing on a special and important appliance – an EV charger.
Detection and identification of sessions of Electric Vehicle charging will unlock loads of benefits for the consumers and the grid utilities. This is of paramount importance since we are adopting EVs globally at an exponential rate and eventually the grids will be under tremendous pressure to accommodate this rapid change. We at Plexflo are solving this problem head-on to help utilities upgrade the grid efficiently for the coming decades.
Analyzing the load profiles and generating analytical behavior of consumption patterns can straight away help manage the grid load. In addition to that, we can also evaluate energy efficiency, recommend better practices for energy usage personalizable for each household, diagnostics, and preventive maintenances, replace expensive submetering, and many more exciting applications.
datastream comes with a robust and efficient Deep Learning model that detects occurrences of EVs being charged. The model can accurately detect and label the timestamps at which an EV might have been charging just by looking at the aggregated grid consumption values (in kWh). We are working hard every day to improve the robustness of this model so that the community can freely work on building prototypes and testing out their hypothesis before adopting our technology towards a stress-free EV adoption. All this contributes toward a greater cause – a “clean and better future for our planet.”
Head over to https://docs.plexflo.com/docs/datastream for more details on how to get started with Plexflo’s datastream and learn about exciting features we are shipping with the Deep Learning model. datastreamis just the beginning of more things to come from Plexflo. Stay tuned and follow us for more updates!