ePrints: Personal Energy Footprinting

Introduction

While buildings are becoming smarter with increasing number of energy monitoring endpoints, the real-time effect of an occupant’s personal actions on the overall energy consumption of the building is still unclear. Building on top of existing indoor localization and building energy monitoring techniques, we propose a system that tracks each individual’s energy usage in a shared environment, and provides visibility into his or her real-time energy footprint. Our system enables accountability of energy usage in commercial buildings, and provides actionable feedback, as well as historical insights for both occupants and building managers, so that they can act accordingly in a timely manner. The system goes beyond traditional disaggregation or monitoring solutions by providing real-time footprint of the users which helps them understand the energy impact of their actions. As per our knowledge, we are the first to develop and implement such a system that is capable of providing actionable feedback in terms of real-time usage footprints for every user under a energy network across different levels of disaggregation – types of loads, location, person, time.

Project Website

The official ePrints website can be accessed here. You can also download the mobile app for iOS and Android.

Implementation

We implemented the system as 5 coupled subsystems each of which designed to take care of specific actions. The Energy Reporting subsystem involves controlling and managing all deployed sensors, plug monitors and smart devices to retrieve all energy actions efficiently, and polling data from the building’s internal building management system (BMS) through the BACNET interface. The Query subsystem takes care of all database and data-structure functions which includes storing data and retriving it for various visualization platforms. The occupant tracking system localizes each occupant based on Wi-Fi fingerprinting and Bluetooth Beacon based localization methods. The energy usage of a user is aggregated based on his location (which room he is at and what are the various devices he is using within the room/building). The database system is an implementation of a custom tripartite graph data structure to handle all related operations with an optimized trade-off between performance and storage space.  The foot printing service is the central service which takes care of coupling all other subsystems and managing the interactions between them.

Another major contribution of our implementation is the development of an Interoperability Adaptor which we call the Interoperability Standardization Adaptor which abstracts all complexity associated with the disparate technology and brands of sensors and commercially available smart devices and monitors. This allows the system to interface with a wide variety of existing smart devices or monitors with limited applications without sacrificing on functionalities.

The implementation was done using Microsoft Azure, Python scripts and various other APIs. We also developed custom sensing nodes on Particle Photon and used existing smart technologies such as Samsung Smart Hub, Aeon Labs Plug load meters, Keen Home smart vents, Cree smart bulbs and other commercially available solutions.

A test system was deployed in our laboratory building and successfully evaluated. We also built web-based visualization platforms in Javascript, and mobile device applications for iOS and Android.

Current and Future Work

We are currently working on scaling the solution for large scale deployment across the entire University Campus or a large scale building like an office or mall. We also aim to include the entire energy network of a user – home, travel, work etc in order to provide a more accurate and complete system. There are several crucial innovations to be dealt with in such a scalable implementation which might lead to interesting explorations into using Wearable Technology, Smart sensing and Connected Devices.

 

Publications

Poster accepted in ACM/IEEE International Conference on Information Processing in Sensor Networks 2016

Best Poster Award, ACM International Conference on Systems for Energy-Efficient Built Environments 2016

Full Paper accepted to ACM International Conference on Systems for Energy-Efficient Built Environments 2017

Snapshots

affect_other_example_bak combined_vis comparison

Contributors:

Peter Wei, Xiaoqi (Danny) Chen, Jordan Vega, Stephen XiaRishikanth Chandrasekaran, Aman Shankar, Nachiket Paranjape, Richa Netto