The goal of our work is to study the role of non-cash incentives in motivating behavioural change. We are motivated by the fact that the rapid adoption of smartphones has made it possible for home users to gain unprecedented access to their home energy usage. Moreover, it is already possible for a smartphone to display home and building energy usage status and trends, and to turn appliances and other equipment on and off.
In this project, we wish to take this work several steps further. Specifically, we want to study what incentivizes a home or building owner to make this behavioural change? Our hypothesis is that the use of social media and smartphones makes it possible to dramatically improve the role of non-cash incentives. Specifically, we would like to use game theory to study the use of peer pressure and competition to drive down energy usage, similar to the use of peer pressure to enforce good behaviour in peer-to-peer networks. We will build social network applications on smartphones that not only give a home or building owner hints on how to reduce energy usage but also incentivize this behaviour.The key focus of our work will be to design a flexible architecture to allow us to experiment with a variety of schemes for non-cash incentives, implementing the appropriate communication and control protocols as necessary.
- A. Pat, K. Larson, and S. Keshav, Big-Data Mechanisms and Energy-Policy Design, Proc. AAAI 2016.
Cloud-oriented architecture for smart home data management
Smart grid initiatives look to generate massive amounts of data concerning consumers’ energy consumption, which although useful is privacy sensitive in nature. This project develops a system to allow third party application development targeted towards this data while retaining its privacy, ownership and control.
- R. P. Singh, S. Keshav, and T. Brecht. A Cloud-Based Consumer-Centric Architecture for Energy Data Analytics, Proc. ACM e-Energy, May 2013.
Energy-optimal routing for smart home sensor networks
In the future, we expect that day to day objects will be enhanced with computation and communication to make them into smart objects. Many smart object networks however, use unreliable communication media such as low-power wireless communication and power line communication, where the communication takes place over the wires in the electrical grid. They also operate in a highly constrained environment in terms of physical size, available memory, CPU power and battery life. The unreliable link characteristics and scarce resources have motivated the design of routing protocols for smart object networks such as the RPL (Routing Protocol for Low-power and Lossy Networks) protocol. RPL is the IETF (lnternet Engineering Task Force) standard IPv6 routing protocol for smart object networks. Although RPL was deemed ready in 2011, it has several undefined specifications which greatly impact its performance. In this work, we have come up with energy-aware routing metrics as well as transmission power control algorithms that decrease the energy cost of RPL.
- Rezaei, E. (2014), Energy Efficient RPL Routing Protocol in Smart Buildings, MMath thesis, University of Waterloo.
Home peak load prediction
Proposed smart grid solutions to reducing peak electricity demand include storage and demand response which require a dynamic prediction of electricity demand. This project involves design of prediction models to predict peak electricity demand of a home on a relatively short timescale such as an hour.
- R. P. Singh, P. X. Gao, D. J. Lizotte. On Hourly Peak Load Prediction, Proc. IEEE SmartGridComm, November 2012. Winner of Best Paper Award in its track.
In our previous study we have observed significant gains due to elasticity of appliances. In this project, we extend the notion of appliance elasticity by proposing a scheme for appliances to respond to signals received from the power grid and a scheme for the utility to compute the signals to send to appliances so that the carbon footprint of the power grid, caused by commissioning plants with high carbon emissions, will be reduced.
- P. Srikantha, S. Keshav, C. Rosenberg. Distributed Control for Reducing Carbon Footprint in the Residential Sector, Proc. IEEE SmartGridComm, November 2012. Winner of Best Paper Award in its track.
Temperature setpoint market
The electrical grid is designed to meet peak loads, which may occur for only a few hours each year. Consequently, there are signiﬁcant economic gains from a reduction in the peak load. Air conditioner (AC) load from residential buildings forms a signiﬁcant portion of peak summer loads. The existing ‘peaksaver’ program in Ontario attempts to reduce AC loads by setting thermostats a few degrees higher in volunteer households on hot summer days. This has had only a limited success. To address this issue, we propose a scheme that provides monetary incentive for participation. We describe the operation of this ‘temperature market’ and demonstrate its effectiveness with a heterogeneous population of potential participants.
- S. Singla, S. Keshav. Demand Response through a Temperature Setpoint Market in Ontario, Proc. IEEE SmartGridComm, November 2012.
Real time data monitoring and anomaly of solar panels
In this project, we monitored, in real time, individual solar panels on the roof of the Environment 3 (EV3) building at the University of Waterloo. This project also comprised design and implementation of hardware and software for detecting the anomalies on solar panels, such as shadows, snow, hail, dust and clouds. A sensor network to measure solar panel outputs was built, using the Zigbee protocol. We deployed the system on roof-top solar panels and developed machine learning techniques for automated analysis and detection of anomalies in current and voltage outputs.
- Hu, B. (2012). Solar Panel Anomaly Detection and Classification, MMath thesis, University of Waterloo.
- X. Gao, L. Golab, S. Keshav, What’s wrong with my solar panels: a data-driven approach, Proc. Workshop on Energy Data Management, March 2015, pp. 86-93.
The effectiveness of pricing on residential electrical storage for the smart grid
Storage batteries that can store energy during times of low load and be discharged during times of peak load, is one proposed solution to reduce peak load. This project evaluates the effect of storage on the electrical grid under different consumer electricity pricing schemes.
- T. Carpenter, S. Singla, P. Azimzadeh, and S. Keshav. The Impact of Electricity Pricing Schemes on Storage Integration In Ontario, Proc. ACM e-Energy, May 2012.
Elasticity of Domestic Appliances
Power demand from the residential sector represents a significant fraction of overall demand. Most schemes attempting to curb residential demand rely upon human reactions to pricing signals. We have discovered an inherent property called elasticity present in most appliances that will allow elastic components in appliances to instantaneously reduce their power consumption without any impact on their lifetime. When these components reduce their power consumption, the time required to complete their operation will increase in proportion. The more the time is extended, the more is the degree of possible reduction of comfort for consumers. We had conducted a preliminary assessment of the peak demand reduction potential for regions with different demographics with various degrees of comfort reduction. This study had indicated significant gains in peak demand reduction even for a minor degree of comfort reduction.
- P. Srikantha, C. Rosenberg, and S. Keshav. An Analysis of Peak Demand Reductions due to Elasticity of Domestic Appliances, Proc. ACM e-Energy, May 2012.
We deployed a testbed of 24 wired sensors in our building and gathered data in sound and light levels, temperature, and wind velocity. We used Partially-observable Markov decision processes to optimally control heating vents to turn off heating and cooling when users are inactive, and to turn it back on when we can predict impending activity based on past patterns of activity.
- O. Ardakanian and S. Keshav, Using Decision Making to Improve Energy Efficiency of Buildings. Proc. POMDP Practitioners Workshop, May 2010.
Smartphone based energy consumption feedback
This project investigated the effect of native smartphone applications on consumers’ energy aware behavior, in light of the now defunct, browser-based commercial services like Google Powermeter and Microsoft Hohm.
Interactive home monitoring and control
Our vision of the future home is of an autonomous power generation system that is capable of modulating its consumption in response to system signals as well as augmenting its energy consumption with renewable resources. This includes the real-time monitoring and control of renewable power generation, power storage, and measurement devices. Designed and implemented a fully interactive energy monitoring and control system for a home, comprising off-the-shelf hardware, existing home appliances, Android smartphone applications, intelligent e-mail bots, and a home energy gateway implemented using Microsoft Home OS. Watch this demo for an abnormal consumption e-mail alert application.