Edge & Swarm: Critical components of tomorrow’s smart grid
“Edge will eat the Cloud,” we’re told. Faced with vast numbers of Internet of Things (IoT) devices producing incredible volumes of IoT data, generating companies are already seriously considering the Edge computing model. But can we point out practical cases where the value of Edge – in combination with Swarm – exceeds that of ‘cloud only’ or centrally-based architectures?
This article walks through a concrete example in the Utilities market, exploring how ‘Edge & Swarm’ offers a scalable option for managing the high penetration rate of distributed renewable energy resources into an electrical grid – in other words, how Edge & Swarm will play an essential role in future smart grids. Let’s start by taking a quick look at Edge & Swarm.
Why we need Edge & Swarm
The Cloud computing model locates the compute resources needed to process IoT data centrally. The incredible volumes of IoT data must be transmitted across networks to be processed and analyzed. The ultra-low latency required is far from guaranteed.
Edge computing offers the answer. Putting compute capabilities very close to the operational field asset reduces not only bandwidth requirements but also latency. The connected edge devices – with their small physical footprints, low energy consumption and intrinsic computing resources – mark a shift away from a reliance on centralized data centers in the Cloud.
To unleash their full capabilities, Edge nodes need to interact in a dynamic ecosystem of connected edge devices, edge compute capability and cloud-hosted services. Orchestration of this collective capability gives rise to dynamic computing swarms able to deliver context-driven services. This is Swarm computing.
Swarm computing offers a new approach for deploying a continuum of collaborative IoT and cloud strategies uniquely capable of dynamically handling even the most complex of scenarios.
Renewable energy sources join the grid
Our example focuses on a future electric grid where massive numbers of distributed solar photovoltaic panels (PVs for short) are being introduced along with new storage capacities, including batteries and electric vehicle charging stations. This distributed production topology risks production peaks that compromise both the balance of the grid and the power quality, as we shall see.
Today’s grid is managed centrally by a system implementing the conventional paradigm of a top-down topology and unidirectional power flow, from large power production units to consumer units. At peak production, local failback mechanisms prompt distributed production nodes to automatically go off-grid to preserve balance for the rest of the distribution network. Corresponding energy production is unfortunately lost as a direct consequence.
This management strategy is not viable in a grid with a relatively high ratio of distributed energy production. So, how do we best implement a smarter grid management strategy capable of:
- Forecasting local production peaks at every grid node
- Querying the grid for available energy storage capacity and
- Reconfiguring the grid before it becomes unbalanced.
Edge & Swarm critical to the smart grid
We could stick with the present mode of operation. While technically feasible, grid nodes – both production and storage – would need to be monitored heavily. Potential drawbacks include:
- Significant data and processing overhead for the central system
- Investment required to connect and instrument additional grid nodes (including low voltage production injection nodes currently typically not visible to the central system) and
- Resiliency issues from OT/IT communication failures.
A mode of operation based on the Edge & Swarm paradigm offers a better option, potentially using so-called ‘smart nodes’ software akin to the Codex Smart Edge solution. Here a mesh of smart nodes software is deployed on fit-to-cost ruggedized hardware at each relevant production and storage node. These now smart nodes host an ad-hoc local production forecast algorithm for forecasting local production peaks.
Visibility of potential protection peaks means nodes can work autonomously overcome resiliency issues. If a node senses it may become disconnected, it proactively communicates with nearby edge nodes to seek storage capacity. After consolidating details of current storage capacity received from its peers, the requesting node conveys the optimal grid reconfiguration option for absorbing the upcoming production peak to the central system. Benefits of this approach include:
- No data or processing overhead for the central system
- Better resiliency since most communication takes place at the edge, and only when necessary
- Ability to deploy ad-hoc local algorithms at the edge for better forecast accuracy.
Overcoming digital dilemmas
A solution based on Edge & Swarm is a safe and scalable option that operators can deploy as an extension to their existing central system to handle increasingly distributed energy resources. This particular use case is discussed in our latest Journey 2022 ‘Resolving Digital Dilemmas’ report.
The report also discusses strategies for addressing the digital dilemmas facing organizations exploring the opportunities brought by Edge & Swarm. In particular, it looks at how establishing and maintaining trust and security within such heterogeneous and dynamic environments will be a significant challenge that demands new approaches, including:
- Distributed ledger technologies could inspire new trust mechanisms
- P2P mesh architecture is an ideal model for defining communication requirements
- Data-ridden security offers a new security paradigm when data is stored virtually everywhere
- Information-centric network looks promising for handling network failures and serious overloads.
Read 'Journey 2022 ‘Resolving Digital Dilemmas', researched and written by the Atos Scientific Community, for guidance on developing strategies to resolve this and other E&U emerging Digital Dilemmas.