Wireless Power Monitoring at Plugs and Panels

Michael Lorek

EECS Department
University of California, Berkeley
Technical Report No. UCB/EECS-2017-176
December 1, 2017

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2017/EECS-2017-176.pdf

In 2012, electricity generation was responsible for over 30% of carbon emissions in the US - surpassing the transportation sector - and predictions to 2040 show this trend continuing with current technologies. Electrical submetering provides improved spatial and temporal resolution into how buildings use their energy, and case studies have shown that improvements driven by submetering data can lead to 5-30% reductions in electrical energy usage. However, traditional building submetering technologies present unfavorable cost, installation, and form factor attributes that inhibit the widespread deployment of these systems. This dissertation details the design and characterization of easy-to-install, low-cost wireless sensors for submetering building electrical power at the circuit breaker and plug load levels.

Discussed first is a sensor installed on the external face of a circuit breaker that non-intrusively measures line voltage and circuit current waveforms, and calculates the real power dissipated in that circuit. The PCB-based sensor uses a Hall effect sensor, a 1920 Hz sampling rate to handle non-linear loads, can measure power levels below 10W, uses all off-the-shelf components with a BOM under $10, and can be installed without hiring an electrician. The total installed cost to submeter an entire panel using the sensor system is roughly $250 - a 10x reduction compared to traditional technologies. Data is presented that verifies the efficacy of the submetering sensor system in a lab setting as well as a real-world residential installation.

Next, the details of a plug-through energy monitor for ubiquitous electrical monitoring of plug load devices in buildings is presented. Using a non-invasive inductive current sensing technique, the current flowing through a plug load device is measured without a series-sensing element that breaks the circuit. This enables slim profile hardware, and eliminates the series resistor power dissipation inherent in traditional current sensing implementations. The prototype can be embedded into an outlet faceplate and easily retrofit onto any existing outlet for long-term measurement of AC power parameters. The sensor includes 802.15.4 wireless connectivity, a 1920 Hz sampling rate, and a measured noise floor of approximately 2W with a BOM around $5.

The new sensor technologies presented in this dissertation are shown to be effective power meters, and are also cheap to build with standard printed circuit board manufacturing processes. Thus, these meters are commercially viable and have the potential to bring building submetering to the masses. Commercial building owners could save over $0.20/ft2-yr by spending approximately $0.15/ft2 to install plug and panel metering. Given an average household, homeowners could save over $130/yr with a ~$375 investment to meter their panel and wall outlets. If this technology was installed in all US residential and commercial buildings, the reduction of greenhouse gas emissions would be equivalent to that of over 2 trillion miles driven every year in a 30 mpg car. In buildings that are very inefficient, a much greater reduction in consumption is possible. With continued development and integration, ideas presented in this work could lower the cost of the metering hardware by another 5-10x, and make the goal of ubiquitous electricity submetering more easily attainable.

Advisor: Kristofer Pister


BibTeX citation:

@phdthesis{Lorek:EECS-2017-176,
    Author = {Lorek, Michael},
    Title = {Wireless Power Monitoring at Plugs and Panels},
    School = {EECS Department, University of California, Berkeley},
    Year = {2017},
    Month = {Dec},
    URL = {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2017/EECS-2017-176.html},
    Number = {UCB/EECS-2017-176},
    Abstract = {In 2012, electricity generation was responsible for over 30% of carbon emissions in the US - surpassing the transportation sector - and predictions to 2040 show this trend continuing with current technologies.  Electrical submetering provides improved spatial and temporal resolution into how buildings use their energy, and case studies have shown that improvements driven by submetering data can lead to 5-30% reductions in electrical energy usage.  However, traditional building submetering technologies present unfavorable cost, installation, and form factor attributes that inhibit the widespread deployment of these systems. This dissertation details the design and characterization of easy-to-install, low-cost wireless sensors for submetering building electrical power at the circuit breaker and plug load levels. 

Discussed first is a sensor installed on the external face of a circuit breaker that non-intrusively measures line voltage and circuit current waveforms, and calculates the real power dissipated in that circuit.  The PCB-based sensor uses a Hall effect sensor, a 1920 Hz sampling rate to handle non-linear loads, can measure power levels below 10W, uses all off-the-shelf components with a BOM under $10, and can be installed without hiring an electrician.  The total installed cost to submeter an entire panel using the sensor system is roughly $250 - a 10x reduction compared to traditional technologies. Data is presented that verifies the efficacy of the submetering sensor system in a lab setting as well as a real-world residential installation. 

Next, the details of a plug-through energy monitor for ubiquitous electrical monitoring of plug load devices in buildings is presented. Using a non-invasive inductive current sensing technique, the current flowing through a plug load device is measured without a series-sensing element that breaks the circuit. This enables slim profile hardware, and eliminates the series resistor power dissipation inherent in traditional current sensing implementations. The prototype can be embedded into an outlet faceplate and easily retrofit onto any existing outlet for long-term measurement of AC power parameters. The sensor includes 802.15.4 wireless connectivity, a 1920 Hz sampling rate, and a measured noise floor of approximately 2W with a BOM around $5.

The new sensor technologies presented in this dissertation are shown to be effective power meters, and are also cheap to build with standard printed circuit board manufacturing processes.  Thus, these meters are commercially viable and have the potential to bring building submetering to the masses.  Commercial building owners could save over $0.20/ft2-yr by spending approximately $0.15/ft2 to install plug and panel metering.  Given an average household, homeowners could save over $130/yr with a ~$375 investment to meter their panel and wall outlets.  If this technology was installed in all US residential and commercial buildings, the reduction of greenhouse gas emissions would be equivalent to that of over 2 trillion miles driven every year in a 30 mpg car.  In buildings that are very inefficient, a much greater reduction in consumption is possible.  With continued development and integration, ideas presented in this work could lower the cost of the metering hardware by another 5-10x, and make the goal of ubiquitous electricity submetering more easily attainable.}
}

EndNote citation:

%0 Thesis
%A Lorek, Michael
%T Wireless Power Monitoring at Plugs and Panels
%I EECS Department, University of California, Berkeley
%D 2017
%8 December 1
%@ UCB/EECS-2017-176
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2017/EECS-2017-176.html
%F Lorek:EECS-2017-176