Edge Analytics in Energy & Utilities Industry
- Dec 8, 2020 7:10 am GMT
As we all know, the edge analytics or we can say AI@Edge is picking up the momentum to bring the computing power at the edge. This is more prominent in Telco industries and have not seen much traction in E&U industry. However. if we can identify the right use cases then Edge analytics could be much more beneficial for utility companies. In fact, if we see the operational technology deployment then the edge analytics is more relevant for E&U industry. e.g. consider the wind farm installations, these are all remote where there might be internet connectivity or bandwidth issues. Connecting wind turbines to cloud or to on-prem data center is feasible but having real time actions flowing from cloud/data center to the wind turbines might be difficult. Another example is the substations, can we have the real time system to avoid physical security attacks? What if the attacker cuts the internet/phone cable first before causing the damages to the transformers? How can we analyze the network traffic anomalies and identify cyber security threats at the substation without reaching out the data center? These are few examples which could lead to identify the right use cases for edge implementation in E&U.
I won’t discuss more about what is edge analytics as one can easily search some literature around it over the internet, but here’s the simplest and short architecture overview diagram for an edge implementation
Layer 1 - IoT + Sensors: This layer includes IoT devices (sensors, smart meters, wind turbines, smart plugs, etc.) as well as users. The first layer is responsible for the ingestion of data and the operations involved.
Layer 2 - Edge Nodes: The second layer is formed by Edge nodes/devices. These nodes are responsible of data processing, routing and computing operations.
Layer 3 - Cloud Services: This layer is formed by multiple cloud services with higher computational requirements. This layer is responsible for Data Analytics, Artificial Intelligence, Machine Learning, or visualization, among other tasks.
Now the big question, what would be the potential use cases and are those going to provide good ROI? Below are few potential use case examples where the edge solutions can be developed for.
Substation Physical Security: Physical security of CIP devices is highly critical and falls under CIP 014 as per NERC CIP standards. If we see the history, there were arm attacks on substations causing huge financial impact for utility companies due to damages at substation and resulting electricity outages for customers. Though we have the security systems in place, but we need a quick and real time actions to avoid such attacks. What if we could implement a video analytics solution at edge which could detect the armed person or loitering around the substation and trigger the real time actions e.g. raising the alarms or activating audio talk down or activating alternative security system.
This solution is highly feasible to develop and implement with advancement in video analytics, camera devices and the edge devices. And the financial benefits/ROI could be multifold considering the loss that could be caused by armed attacks.
Why the edge computing solution for this use cases and not a cloud based? Two main reasons:
Network performance/latency – Transferring the huge amount of video data to the data center to perform the analytics may cause delays in the response time for action to be taken. The analytics can be moved to the edge which can result into quick action before the damage is done.
Network Failure - There might network problem and the action data may not come back to substation on time.
Substation Logical Security: Another use case around the substation is the cyber security of substation devices. We can have an edge solution to analyze the network traffic and identify the anomalies which could potentially cause the cyber-attacks on the substations.
I have mentioned the use case examples which look pretty simple. However, designing and developing the solutions will need good number of efforts which may include combination of AI modeling and deployment over the edge devices. We also need to consider the integration and security aspects, how to have this installed at the substation devices and how to securely allow the communication between electricity devices and the edge devices. E.g., if we want to build an edge computing solution for wind turbines then there are multiple integration questions that need to be answered e.g. how to read the SCADA data into the edge devices and at what frequency, and how to securely connect edge devices to wind turbine PLCs etc. The overall installation and maintenance cost also need be considered while designing the solution. So, it’s a joint effort which may include the E&U SMEs, client, business analytics, data scientists and IOT developers. But yes, if designed and implemented with the right strategy and roadmap, this technology is going to be highly beneficial for utility companies!
Get Published - Build a Following
The Energy Central Power Industry Network is based on one core idea - power industry professionals helping each other and advancing the industry by sharing and learning from each other.
If you have an experience or insight to share or have learned something from a conference or seminar, your peers and colleagues on Energy Central want to hear about it. It's also easy to share a link to an article you've liked or an industry resource that you think would be helpful.