Smart Energy Transition and Edge Computing

Digital technology, pervasively, is getting embedded in every place: every thing, every person, every walk of life is being fundamentally shaped by digital technology — it is happening in our homes, our work, our places of entertainment. It’s amazing to think of a world as a computer. I think that’s the right metaphor for us as we go forward. — Satya Nadella, CEO of Microsoft

The internet is no longer a web that we connect to. Instead, it’s a computerized, networked, and interconnected world that we live in. This is the future, and what we’re calling the Internet of Things. — Bruce Schneier, security technologist, and author

With the world undergoing an energy transition driven by the perils of climate change, this paper discusses how developments in edge computing through the internet of things could drive innovations to create a circular economy.

As per a survey of worldwide telecommunications professionals, the Utilities and Industrial IoT show great market potential. In addition, over 70 percent of respondents worldwide agree that the use of smart technologies in the home makes life easier. Smart devices are a type of IoT enabler for edge computing to facilitate the energy transition.

Some of the applications and benefits of IoT in the energy sector include:

• Demand response for residential, commercial &industry users reducing demand at peak time; which itself reduces the grid congestion and electricity bills and the need for investment in grid backup capacity
• Smart building for centralised and remote control of applications and devices improving readiness to join a smart grid or virtual power plan and improving the integration of distributed generation and storage systems
• Battery energy management leading to data analytics for activating battery at the most suitable time reducing the cost of overall energy use

Various applications of IoT in the Energy Utilities industry

Retail electricity customers appear to be developing an overall positive attitude towards smart devices, as evidenced by this relatively recent survey of Finnish energy consumers.

The rest of this paper presents an architectural pattern of how smart/ home IoT devices could leverage technological developments in real-time event streaming and enablers such as 5G and FTTP to construct an accurate picture of energy utilisation at homes. Energy companies can then apply the insight for both supply side as well as demand-side management. For example, drive efficiencies in energy trading, and consumption.

For illustration purposes, the discussion below borrows illustration from Google’s IoT Core product, which is a fully managed service that allows a user to easily and securely connect, manage, and ingest data from millions of globally dispersed devices. Energyworx for example has used the IoT Core platform to build an energy data management platform that enables behavioral science and lets businesses aggregate data in several profiles with valuable context-enriched data points.GCP’s Reference Architecture for IoT Core

Fig 1. Google IoT Reference Architecture

Cloud IoT Edge​ is a set of software packages and services that turns Linux based devices into full-featured IoT edge devices capable of running and applying ML models at the data source. Devices can be anything for many different use cases — from patient-centric medical devices to robotic arms, oil rigs, wind turbines, and the like — any device for which it makes sense to apply machine learning in real-time, onboard the device. At the hardware level, Cloud IoT Edge devices can include one or more Edge TPUs (tensor processing unit​). An Edge TPU is an extremely small hardware accelerator ASIC(application-specific integrated circuit)developed by Google for the TensorFlow ​framework. Open-sourced by Google, the TensorFlow API and framework are used for machine learning applications and neural networks. (Cloud ML Engine — ​ the serverless, a fully managed core component of the Google Cloud Platform that is the centerpiece of Google’s machine learning capability for GCP — can be thought of as hosted TensorFlow.) Several Edge TPU chips can fit within the circumference of a penny, so even the smallest devices can be equipped with machine learning capabilities. Read more about it here.

Non Invasive load monitoring (NILM)techniques are traditionally used to disaggregate energy consumption within a property using transient and steady-state device signatures (Fig 2).

Fig2. Transient and Steady-State noise signatures used in NILM techniques for load disaggregation

However, it isn’t as effective where the devices have complex states like washing machines, dishwashers. One reason for that is that those devices change states more frequently than the usage data (for example half-hourly) captured by the (smart) meters. For example, refer to the illustration below.

Fig 3a. Device signatures

Fig 3b. Half hourly smart meter data profile (illustration only)

Here we discuss how smart IoT capabilities could complement the process by enabling more accurate device state capture in near real-time.

In this reference architecture, devices would publish their telemetry state information (eg what state of the cycle a dishwasher or a washing machine is) in real-time to a topic (Fig. 3).

Fig 3. A simple example of IoT data flow between edge devices and a server

Fig 4. Device configuration and Telemetry

Data analytics techniques could then

1. Join the device state information with industry-wide NILM databases of the device signatures and device fingerprints to more accurately infer its usage energy usage data. Google Cloud IoT Core enables end devices to be configured centrally through a device manager across the internet.
2. Superimpose the disaggregated energy information to smart meter energy data to generate insights into energy usage habits of households and industries

Energy companies can recommend energy-saving tips and promote incentives to influence energy consumption habits. Monitoring energy consumption and other parameters are estimated to allow for a 40% reduction in energy consumption.

A few themes to innovate on top of an IoT enabled energy data management system

1. Household device usage rule engine: Households could set rules defining when specific devices could be used and for how long. They could remotely configure usage patterns of light bulbs, fans, televisions, etc for example when they are away from the property
2. Device Hub: A digital control panel for household and commercial users easy access to their device configuration — for example usage statistics, the ability to set themes (such as lighting, etc)by rooms
3. My Offers: Presenting customers Offers (eg cashback, etc) if certain target usage patters are met (eg not using washing machines and dishwashers during peak demand period)

Thank you for your time and if you would like to add to the discussion, please feel free to leave your comments or connect with me over LinkedIn.

Other References:

1. Building a smart home cloud service: https://medium.com/google-developers/building-a-smart-home-cloud-service-with-google-1ee436ac5a03

(PDF) Intelligent Edge Computing for IoT-Based Energy Management in Smart Cities

2.https://www.researchgate.net/publication/338684011_Internet_of_Things_IoT_and_the_Energy_Sector/link/5e244ecc299bf1e1fabdce64/download

3. https://www.youtube.com/watch?v=X_T8KDQLe2Q