Paper Discussion: 3b. From Energy Audits to Monitoring Megawatt Loads: A Flexible and Deployable Power Metering System

[AkinoriKahata]

• @searri, Rick searri, Critical: In my understand, gateway is a kind of edge computer, then I think a strage of a gateway can be enough size. But how much devices can be connected with one gateway depends on spectrum problem, then it is limited.
• @RyanFisk2, Ryan Fisk, Comprehensive: The paper did not shows how install practically, but it show Triumvi can be used for many opportunities. Major limitation of the meter is tradeoff between accurate sensing and flexibility, and this paper shows flexible sensor device can measure accurate enough. Real deployment problem is next challenge.
• @chandaweia, Cuidi Wei, Critical: From the sensing device structure, I guess Triumvi device may not cause any damage for systems. But I think stability problem is there, this system have to be experimented under real environment.
• @anguyen0204, Andrew Nguyen, Comprehensive: Whole system cost is not calculated. These system needs deployment cost and managing and maintenance cost, then the total cost will be appeared after more real experiment. It`s next challenge.
• @ericwendt, Eric Wendt, Comprehensive: I think this system can not detect small current, then gateway should distinguish valid data or not when it receive the packet that shows current is small enough from the devices.
• @huachuan, Cuidi Wei, Critical: In my understand, data storaged in FRAM and data storaged in gateway are different format. Because meter can not have enough computer resources and electric power, then it can contain only raw data. On the other hand, gateway can storage after data decoupling.
• @zacharied, Zack Day, Critical: In the paper, only encryption is relevant to security, it is most important thing for IoT sensor device though. Gateway has a enough resources, then we install more security thing for gateway, but the difficulty is limitation of sensor devices. How to fit the real world is next challenge.

[searri]

Reviewer: Rick Sear Review Type: Critical

Problem being solved

We really don't have an effective but cost-efficient way to meter power use accurately. Other problems with existing tech are that it can't be installed, since it requires downtime which can't always be achieved, and also the existing tech is difficult to install, even with the power off. To solve this, the paper proposes Triumvi, a power monitor that can be easily installed without turning off the main power.

Important areas

The paper details a solution to the current issues with power monitoring by proposing a device that is easy to set up without requiring power to be off. The Triumvir device also has different modular upgrades to make it usable in different areas. The device was pretty extensively tested to ensure these bold claims held up.

Questions

• If I tamper with the Triumvir device so that it charges an arbitrary power consumer rather than itself, couldn't I harvest lots of unmetered power? How much power does this device harvest (i.e. is my "free power" scheme even feasible), and is the device secure enough to prevent this from happening?
• Section 4 talks about the different modular upgrades for the device. While this is cool, how much extra complexity does this add? Is there a point where I'd need to hire a specialist to do it for me?
• Section 5A (implementation of the power harvest/measurement switching device) went way over my head, but I'd really like to understand it. It seems that the device can switch modes without any computing? How durable/safe is this setup?

Critiques

• The Triumvir device needs a gateway device to function properly. The paper does not mention if this is a 1:1 ratio (I assume not), but even so, isn't this adding extra hardware (cost, power consumption) and installation complexity to a device being billed as very simple and easy to use?
• The paper doesn't mention (or didn't test) what happens if the network gateway is unavailable. How much onboard storage is there? Does data get lost or overwritten? The absence of the gateway, even temporarily, might really harm this device's usefulness.
• They did a good job testing the power measurement accuracy and use cases (obviously most important). Since a big selling point of this is how easy it is to use, I would also have liked to see more extensive measurements on how long it takes to install for new users, especially considering all the possible upgrades for the device.

[RyanFisk2]

Reviewer: Ryan Fisk Review Type: Comprehension

Problem being solved Being able to monitor power usage at a circuit level could not only cut on on the cost of electricity, but also allow the electrical system to more efficiently allocate power to areas that need them with a high degree of specificity. However, most of the modern power meters won't work at a large scale because they are too big, require their own power source, or require an electrician to set them up. Installing any such meter would require the whole system to be brought offline temporarily, which in some cases isn't feasible.

Paper contributions This paper attempts to solve most if not all of the issues with power metering with Triumvi, a small power meter that clips on to the circuit it measures and siphons power from the system to power itself. Since the device is small and wireless, the paper argues that it would be usable in large scale power metering. Triumvi can also communicate wirelessly with a gateway which would allow power readings to be sent to a cloud for processing.

Questions I still have Exactly how scalable is this system the paper proposes? The paper specifically has a section about commercial and residential buildings, but that's as far as the authors speculate.

The paper mentions that the Triumvi device has a minimum required load for it to operate (since it draws power from the system it monitors). However, it doesn't go the other way. Is there a maximum voltage that this system can handle before breaking down and would it be able to report that a lethal (at least to itself) surge was happening?

How many devices could one gateway serve and how many gateways would you need to effectively monitor a residential or commercial building?

[chandaweia]

Reviewer: Cuidi Wei Review Type: Critique

Problem being solved Existing circuit-level metering systems are too costly to deploy, due to difficult installation or cumbersome calibration processes, too inaccurate, due to an inability to faithfully calculate power from synchronized current and voltage channels, or too unreliable, due to a strong dependence on a frequently lossy wireless channel. So this paper proposes Triumvi to help make circuit-level metering affordable, accurate, and reliable.

Main contributions This system addresses many of the challenges identified with building energy monitoring and builds on prior work from both commercial and research settings, to address the U.S. Department of Energy’s calls for new methods of power metering.

Questions about the paper

1. What about the Triumvi’s measurement rate compared with other meters(For Fig.10)?
2. How about the stability of the Trumvi? Although the Triumvi is self-powered and non-contact, true powering metering system, the stability of Triumvi needs to be measured.

Critiques about the paper

1. Even if the Triumvi indeed can improve the costly deployment. Are there any other disadvantages the Triumvi incur to influence its function?

[anguyen0204]

Reviewer: Andrew Nguyen Review Type: Comprehension

Problem Being Solved: There needs to be a more efficient, cheaper, and practical way to observe electrical usage in buildings down to a circuit level. Normally, there are submeter devices that could perform these tasks but they contain those aforementioned problems. The paper introduces Triumvi and its capabilities in being able to efficiently tackle this problem with its self-powering and cost efficient qualities of being effectively usable.

Contributions: How actually useful and feasible it would be is the question. Although Triumvi may be able to offer a lot, it still begs the question at how long lasting and stable it could be when deployed to different stations and areas. In addition, there are a wide variety of locations and it has to be considered that there are different complications that it could face. The paper delves into this by explaining the various load rates and how it can impact the system and such.

Questions:

1. How exact are the costs for Triumvi? Could it fluctuate and change given the basis of such various factors that can contribute to how optimal and efficient it functions in various spaces? It's currently estimated at $7
2. There seems to be a few different ways of maintenance and with different options of doing so. How safe and effective would it be?

[ericwendt]

Reviewer: Eric Wendt Review Type: comprehension. Problem being solved: Apparently the electricity usage monitoring process is not streamlined, or rather could be streamlined better. Enter Triumvi. Triumvi allows for a simple clip on device to attach directly to power lines for low-cost power monitoring. It also boasts a range of other appliances such as harvesting power through supposed inductive loops and networking to a home area.

Contributions The best part of this product is its ability to clip onto power lines to monitor usage. This is an extremely non-invasive method of gathering information. It also removes the need for an electrician to set this up. As long as the consumer can follow basic directions, they've actually got a decently optimal device here.

Questions:

• The paper mentions that the device will send packets to the gateway if current that was too small to be detected is passing through. How does this affect the error calculations of this device?
• How does the device work with single-phase current?

[huachuan]

Reviewer: Huachuan Wang Review Type: Critique

Overview This paper proposes Triumvi, which is a self-powered metering system that can help make circuit-level metering useable. Triumvi harvests energy to power itself monitors current, and voltage calculates power, encrypts data, and wirelessly transmits the results. Triumvi is easy to install, upgrades, charge sharing, and current waveform analysis.

Contribution This paper pointed out that submeter building electricity usage accurately, inexpensively, and wirelessly submeter, making electricity usage at the circuit-level, is a very critical topic. Triumvi meats the requirement of DOE Low-Cost Wireless Meter Challenge, and offering simple installation procedures, high measurement accuracy, secure data transfer, and affordable energy metering. Triumvi can provide energy data in the industry sector and can benefit energy characterization.

Questions

1. Decoupling data from the wireless radio is confusing. The radio module is for wirelessly transmit their data for collection. Why the radio that is the best for transmitting to the gateway is likely not ideal for getting the power data to a smartphone?

2. This paper mentioned the gateway can offload the collected data to a cloud backend for storage and processing. Why the gateway forwarders batch readings before sending them to the cloud instead of batch sending them directly?

3. Triumvi only extracts the voltage phase to avoid calibration. Because lots of information e.g., the geometry of the breaker, will affect the measurement, how does to prevent the calibration at runtime?

Critiques

1. This paper mentioned that the gateway can log data to an internal database for offline processing. However, it does not indicate the storage capability of the low-energy memory (FRAM) that is for store data locally is.

2. RTC might not necessary because the Triumvi node can send a time request packet to the gateway which immediately responds with the current time.

[zacharied]

Reviewer: Zach Day Review type: Critical

Overview

The paper introduces Triumvi, a system designed to address the U.S. Department of Energy's calls for efficient circuit-level power usage monitoring. Triumvi is one of many designs for power monitoring solutions. Its main focus is on ease of deployment.

Contribution

According to the paper, no previous power metering system has been able to address key deployability concerns. For example, systems with a high accuracy are difficult to deploy, with some even requiring downtime of the circuit as a whole in order to install, which makes installation impossible for many providers. Triumvi solves this problem via simple clip-on meters, attached to circuit breakers, that are able to detect current in the circuit with a high degree of accuracy. The meters are capable of powering themselves with the circuit they attach to, as well as wireless transmit their collected metrics to a cloud service.

Questions

• The paper mentions the non-contact clips as being a major advantage of Triumvi, but does not go into depth on the trade-offs made as a result of a non-contact solution. I would like to know more about the pros and cons of each approach.
• They mention that, in the event of network outage, Triumvi clips can store data in FRAM. The way they describe implies to me that it's intended as a fallback, but how did they decide on the amount of FRAM used by the system based on the trade-off between storage amount and power consumed?

Critiques

• Their analysis of installation overhead (p10) feels hand-wavey; their definition of inexperienced user is vague, and I feel like they could have written any statistics they wanted here, shifting their definition of an experienced user to whatever is most convenient.
• Given that this is an IoT device, I would expect more emphasis to be placed on the security and encryption factors. However, this is treated more as a side-note to them.
• Perhaps a nitpick, but the real-life photos of the hardware used by Triumvi feels less suited for a technical document and more for marketing material. It would be much simpler to just provide the dimensions of each device, if such data was really critical to the paper.