Design Guide ESS.WP.925. Redundant Ethernet 01
Mission Critical Ethernet Networks For Generator Fuel Systems
Introduction
We have used Ethernet Networks for over 10 years as a means of connecting distributed control
panels and providing monitoring information to our Master Panels and the wider BMS and
Building System network. Ethernet is fairly easy to setup, fast, and reliable. Industrial Ethernet
hardware has evolved to meet tough operating environments. And using fiber optic cabling,
distance limitations (100 meters for cat 6 cabling) are usually not an issue. However Ethernet
systems are subject to disruption, and require good engineering for mission critical applications.
Monitoring versus Control Functions
Ethernet networks provide reliable monitoring of fuel system status information. Systems have
also included hard-wired control signals for critical functions, such as a day tank controller
commanding a duplex pump controller to start a pump. A simple Ethernet network using
unmanaged switches is often acceptable where the network function is only monitoring.
As the perception of Ethernet reliability has increased, designers are starting to require fewer
hard wired signals, relying on Ethernet communications for important operating commands.
Ethernet for fuel system control and monitoring requires good design practice for the Ethernet
networks.
Ring Topology
A ring network is a single-fault tolerant ring network intended for the interconnection of devices.
Advantages
-
Ability to survive a single point of failure or a device being powered down on the ring.
-
Simplified cabling
- Ability to cover long distances with 100 m allowed between each copper segment
Disadvantages
-
Additional configuration complexity
-
Longer convergence times
-
Performance can be difficult to predict
As the perception of Ethernet reliability has increased, designers are starting to require fewer hard wired signals, relying on Ethernet communications for important operating commands. Ethernet for fuel system control and monitoring requires good design practice for the Ethernet networks.
Ring Concepts for Network Reliability
Much of the thinking about redundant Ethernet networks is based on ring configurations. A simple explanation of the ring is that a device connects either clockwise or counterclockwise around the ring. A single device failure can be overcome by re-routing the network traffic to that device through managed Ethernet switches designed to perform that function.
Networking at the Front Lines of Power Reliability
One of the concepts implicit in this system is that the devices themselves are highly reliable –
and that the power sources for these devices are highly reliable. Fuel system controllers, in
contrast, are at the front line of that power reliability, where they will likely see more power
disruption. A better than ring networking scheme is needed.
Double Star Network for Multiple Failure Tolerance
A double start network configuration can tolerate multiple faults, and the recovery time for any
fault is much faster than a ring configuration. The double star may require more cabling than a
ring, however it is easier to diagnose and troubleshoot during start up and for future
modifications, because all cable are accessible from the two master panels.
Network Self-Monitoring
Good fuel system networks should include a self-monitoring feature, where connections are
continuously checked, usually via heartbeats. Network failures should be identified and alarmed
both locally at the master controllers and through the BMS / PMS systems.
