TI Training home
Wireless network challenges and solutions for a smarter grid IoT
Introduction Introduction: Smart Grid IoT Overview and Trends
Wireless network challenges and solutions for a smarter grid IoT
1.2 Introduction: Smart Grid IoT Overview and Trends
Hello, everyone. In this training session, we will discuss smart grid IoT overview and the key subsystems for smart grid IoT and then we discuss them on the trend. Let's start with what is IoT for smart grid applications. As many of you know, IoT is defined as connecting things to the internet. I have a color code here to mapping each term to key technologies or metrics.
In terms of connectivity, reliable connectivity and network scalability are very important. Sometimes, we see data rate requirement, but it's very dependent on the applications. Since the connectivity is installed in a thing, typically, it is limited to processing power and memory. It's different from the cell phone or PC installing the connectivity.
In addition, many of devices operate on the battery power and so power consumption is also a big concern. To install connectivity on a thing and to provide the unique identifier to each device, IPv6 providing the system bite address is a must. To support internet services on the thing, there are multiple key technologies to realize this.
Cloud is one of a popular technology platform to serve as a centralized location for data storage and analysis. Stand out TCP/IP protocols are also a must to set to support internet service and 6LoWPAN is recently introduced to reduce the gap between the IP and the lower level stacks to serve internet service on a thing. We will discuss this in details later in this session.
There are multiple target and equipment in smart grid domain. Here, we introduce some equipment, but again, this is not limited. For smart meters, electrical meter and the gas and the water meters and the heat cost allocators are key and equipment for smart grid IoT. Moving to the grid automation, [INAUDIBLE] indicator for monitoring and the remote terminal units are also key equipment for the smart grid IoT. For renewable, wind and solar energy can be a strong candidate for the smart grid IoT, and then the electrical vehicle is also another area to consider smart grid IoT.
Now lets go into a little deeper to understand the wireless networking systems and trend. To realize the smart grid IoT, there are multiple keys of systems needed. The first one is, on the left side, is end node with connectivity. And edge router in the middle play a key role to bridge the end node with the backbone and the cloud in that term.
Typically, the edge router or gateway installs two connectivity solutions to talk to each of the network. Then, the control center, PC, or cell phone that can trigger, monitor, or control activity to the end node where the edge router is needed. These nodes have, typically, the backbone connectivity only. We call the network between the end node and edge router a access network and the network between the edge router and the control center as a backbone or a cloud network.
Let's first take a look at the access network here in the red box. Many different wireless or wired connectivity solutions exist in this domain. Typically, sub-1 gigahertz RF is a strong candidate in this domain because of its propagation property is excellent. In addition, 2.4 gigahertz is also used in many different reasons for connectivity solutions. For topology, star and the mesh topology are mostly used. Here, the table in the right bottom summarizes the pros and cons for each topology.
Sub topology is a low complexity in software because it does not require any routing protocol to form the route. Wire mesh routing, mesh topology, increases that complexity because the mesh routing, of course, the software is needed in as a stack. System mesh routing allows the connecting to device via multiple hops. It provides great flexibility and network scalability in terms of the connectivity. Of course, a star topology does not provide any method if a device can be reached without the link level connection so this is why sub-1 gigahertz RF is popular with the star topology because it can achieve a longer distance reachability compared to the 2.4 gigahertz.
Now let's move on to backbone cloud and network trend. One question here is who will perform data analysis at your router or a cloud server. In a smart grid domain, we actually observe both. The table on the bottom shows comparison for each approach. When edge router performs data analysis for the information collected from end node, it increases the software complexity and the hardware requirement on the edge router. However, due to the local analysis feature, it reduce the network traffic greatly to the backbone and achieve quicker response time to the node.
Especially, we will observe that this trend in a substation application domain where real-time control is very critical in this domain. On the other hand, the cloud server based approach reduced workload in edge router and so it reduced the software complexity and hardware the requirement on the edge router. However, since role information has to be delivered to the cloud server for analysis, it increases network traffic to the backbone. And since it goes through the more passes compared to the edge router approach, the response time is relatively slower than edge router approach. But my personal opinion is since, typically, the backbone communications support very high data rate so I think the delay with additional pass would not be very critical. This will be very dependent on the application specific requirement.
This slide shows how application level connection is set up between a end node and the control center. A trend we observed is that end node is built with IP and application layer for IP and application stacks and so end node directly connected to control center. In this case, the role of edge router is a pretty minimum just acting as a bridge between the control center and the end node. Another trend is that edge router or gateway plays more roles in connecting between end node and the controller center. So end node are completely hidden from control center and each router locally maintained all the end node and provide the necessary control to end node based on the control centers request or deliver information to the control center.
We observed the direct application level connection between end node and the control center with [? co-app ?] based applications and second approach just a local connectivity that has been seen in the web based application where web server is located in the edge router. There are many connectivity solutions in access network, backbone network, even cloud service domains.
In the access network domain, there are many different wireless solutions such as sub-1 gigahertz RF wireless M-Bus, Wi SUN, and 2.4 gigahertz ZigBee and the Wi-Fi and the celluar based approach like Sigfox and LoRa and the Narrowband IoT and LTE-M and so on. A challenging part, as you can see here, is that it's not likely that a single wireless solution will win for the entire IoT.
Now, in the middle moving to the backbone connectivity, it looks like much less solutions where typically the ethernet or ethernet via the Wi-Fi access is mostly popular social solutions in this domain. In the cloud domain, there are multiple vendors for the cloud service. Recently Amazon Web Services get a big spotlight on the cloud service and also the IBM, Apple, Microsoft also have this on the cloud services. OK.
This is all for the smart grid IoT overview and trend training session. At the next session, we will discuss the challenges on the smart grid IoT. Thank you for watching.
2017년 4월 10일
This training will cover system- and software-level deep-dives on key communication protocols which comprise a 6LoWPAN-based RF network. The training will review new TI Designs/products for an RF network implementation on the CC1310 wireless MCU and Beaglebone Black (BBB)– including both the end node and edge router (or data concentrator) solutions.