1 May 2006
Network, Organize Thyself
Industrial environments could benefit from reliability of self-organizing wireless networks
By Gabe Sierra, Dan Carlson, Bob Karschnia, and Brandon Robinson
When process personnel hear the term wireless instruments, they immediately think, great, no wires. But how do you know if you are receiving good data on time, every time? To go a step further, what are the limitations of self-organizing wireless network reliability, and where should you apply it in a dynamic industrial environment? Can wireless field devices ever be reliable enough for today's complicated industrial environments?
Network reliability is the ratio of received messages to sent messages between the gateway and the wireless devices in the network. This is a measure of the end user's ability to receive desired data from field devices. Network latency, associated with network reliability, is the amount of time it takes the message to travel wirelessly from the field device and have the host receive it. Overall network reliability depends on the sensor, wireless communication, host integration, and data management. The complete wireless network can only be as reliable as its weakest component.
Energy conservation and management are at the heart of all wireless protocols. Ultimately, wireless instruments will need to self-power a radio, a sensor, and embedded electronics. If we expect primary batteries to replace wired power, wireless networks will need to operate on duty cycles where all electronics are off (in sleep mode) when they are not sensing or communicating.
Duty cycling a sensor is easy with a simple timing circuit, but what about duty cycling an entire network? Not only will the devices have to wake at the same time, but they will have to be dependable in organizing and relaying messages.
A more reliable network
Star topologies (a limitation of traditional point-to-point and line-of-sight solutions) often have path reliabilities as low as 40%, which is less than ideal. Path reliability is the ratio of received messages to sent messages between two wireless devices in the network or a wireless device to the gateway. For self-organizing networks, the path reliabilities are typically about the same as traditional line-of-sight technologies. But there is a key difference: Self-organizing networks yield network reliabilities in excess of 99% by enabling wireless devices to act as intermediate routers, thereby providing alternate paths (redundancy) when a message is not able to take a particular path. The key is with message redundancy through multiple device neighbors within several hundred feet.
Self-organizing device vendors will differentiate themselves in part on the devices' automatic ability to efficiently decide which path is open. A well-groomed network would show paths AB and AD are more reliable, and AC is a back-up path. This functionality allows networks to quickly self-adjust to the changing industrial environments and produce scalable networks that can easily span entire process units. You don't need expensive, unreliable site surveys, which are only valid for a snapshot in time anyway due to the dynamic environment.
If a truck or scaffolding were blocking path AB, the wireless device would primarily use path AC and AD. The network would identify this on its own and self-configure to optimize the network, during the installation and for long-term operation in a changing environment. This ability creates ease of ownership because it is the only emerging technology that can reach the out-of-the-box ideal state. Even the most novice plant technician can install and operate a self-organizing net- work if it leverages industrial sensor standards and information systems.
Two fundamental methods exist for synchronizing wireless devices for communication: time division multiple access (TDMA) and collision sensing multiple access (CSMA). With TDMA, each set of wireless devices knows exactly when and how often to communicate within the network. In the previous graphic, Device A will transmit to Device B at scheduled times, and Device B will store and forward to another device or direct to the gateway at its scheduled time. The key advantage is each message will have a specific time slot to communicate through the network, thus mitigating issues associated with message collisions. TDMA provides a sturdy way to achieve true scalability in a plant with thousands of devices; it further enables co-existence with other wireless technologies and solutions in the plant. Also, not all devices in this type of network need to be awake at the same time, and this extends device battery life.
Conversely, CSMA requires all devices (not just a single pair of devices) in the network to wake at the same time and try to communicate. If a message from Devices B and C arrive at the gateway at the same time, then they will collide and retry to send the message through a different path. The key disadvantage for CSMA networks is scalability requires higher power consumption due to increased communication requirements from resending collision messages. This imposes adverse implications with device battery life. As the number of wireless devices in a network increases, the number of collisions increases. For a given duty cycle window, where all units are communicating, only a finite number of smart messages can get through before jamming communication paths with retries. So you can't predict network reliability and latency or make anything but the smallest networks viable on primary batteries. If wired sources powered CSMA systems, infinite time for retries (and long latencies) would be available, but it also defeats the advantages of a truly wireless solution.
Frequency agility essential
Another reliability must is to make sure communicating devices have clear frequency channels for successful message transmission. With the proliferation of wireless technologies in a plant, such as radio frequency identification (RFID) systems, it is possible for wireless devices to jam. So you need frequency agility. The most common form of frequency agility is frequency hopping, whereby a pair of neighbors will try new channels until a clear channel allows communication. This enables wireless devices, wireless fidelity (WiFi), RFID, and other wireless technologies to co-exist.
For TDMA networks, knowing which frequency to communicate on is included with the synchronization and advanced network management algorithms. Considering the message transmit-time between neighbors is on the order of milliseconds and the network can operate on many different frequencies, multiple mesh networks can easily co-exist in the same process unit.
For CSMA networks, all wireless devices would have to communicate simultaneously across the available frequencies, which will lower the network reliability due to increased message collisions, longer latencies, and higher device power consumption.
The primary drawback for wireless networks is the power consumption limiting device battery life. Unless a user is willing to replace batteries every few months or hard-wire a power source, measurement update rates will be significantly slower than wired systems to maximize the time between maintenance needs.
The target application for first-generation, self-organizing network solutions will be measurement updates of nearly once every 60 seconds. This is a different paradigm from wired systems, where milliseconds are more the norm. Although not initially viable for high-speed, closed-loop control, secure self-organizing networks with 60-second updates will be well suited for typical monitoring, open-loop control, and some latency-tolerant, closed-loop control applications.
An ideal opportunity for self-organizing networks would include the hundreds, if not thousands, of manual monitoring points. Compared to data collection schedules (update rates) of once per shift, day, week, month, or never, a once-per-minute update rate appears more like real time. Furthermore, it removes clipboard notation errors, inaccuracies from dial gage indication, and poor repeatability of handheld measurement equipment.
Potential open-loop control applications include those in which it might take an operator a while to obtain the appropriate work permit or go out to the site to perform the appropriate control action, such as turning a pump on or opening and closing a manual block valve. In these cases, 1-minute updates are much faster than the time it would take an operator to go out to the relevant location.
The key to managing network latency of self-organizing networks is to ensure the selected network has time stamps at the wireless device for each data sample and provides acknowledgement for successful transmissions. The dependable timestamp will allow compensating network latency. Acknowledging this will verify the message arrives unaltered at the host. As improved power efficiency and management materialize, the ability to increase the measurement and transmission rates will make more non-safety and critical applications feasible with self-organizing networks. Until then, users could incorporate hundreds or even thousands of measurement points into a typical plant. This will empower them to optimize their assets and make them more competitive in the global marketplace.
About the Authors
Brandon Robinson is an instrumentation and electrical (I&E) technician at EnCana in Parachute, Colo. Dan Carlson is development engineer, Bob Karschnia is director of technology, and Gabe Sierra is wireless marketing manager, all at Emerson Process Management in Chanhassen, Minn.
Time division multiple access (TDMA): A method for synchronizing wireless devices in which each set of wireless devices knows when and how often to communicate within the network.
Collision sensing multiple access (CSMA): A method for synchronizing wireless devices that requires all devices (not just a single pair of devices) in the network to wake at the same time and try to communicate.
ZigBee: A published specification set of high-level communication protocols designed to use small, low-power digital radios based on the IEEE 802.15.4 standard for wireless personal area networks (WPANs).
Bluetooth: A global initiative to set a standard for cable-free connectivity between mobile phones and PCs as well as handheld computers, using short-range radio links in the 2.gGHZ Instrumentation Scientific and Medical (ISM) free band.
Water System Unplugs, by Victor K.L. Wong and Thomas Dunn
Wireless SCADA Gains Foothold, by Alison Diana and Hailey Lynne McKeefry
ISA-SP100, Wireless Systems for Automation
Aerospace, Gas Industries Need Wireless
"If you ask someone how to build an airplane, the answer may be informative and useful on a high level. However, if you are planning to make the investment to actually build one, you will not find the one person that has the expertise in all the disciplines needed to produce the end product cost effectively," said Mark McNerney in validation instrumentation at Boeing Commercial Airplane Company in Seattle, Wash.
McNerney said it's the same with wireless standards, which need experts in all associated disciplines working together to make sure details are in line, and it's completed in a time frame that makes it useable. "That's the primary reason standards will continue to be important to Boeing," he said, "interoperability and obsolescence."
One way Boeing is planning to use wireless technology is in the area of flight test, McNerney said. "During a flight test program, the cost of wired sensors, in terms of installation and removal labor and cycle time, is substantial. By creating wireless sensor networks, that cost will be reduced and our testing will be more flexible, adding value to our product." McNerney's team needs equipment they can rely on to "integrate easily into our flight test data systems, operate without extensive development effort, and replace with backward compatible new equipment when parts can no longer be acquired or when new technology drives us to more advanced and capable designs," he said. "We also need to be able to leverage COTS designs to keep costs as low, while responding to emerging requirements."
Boeing sees standards in wireless, such as those emerging from the ISA-SP100 committee, as a "tool in the daunting task of coordinating our wireless applications both in terms of quality of service and interoperability," McNerney said. Although the standards won't replace testing and validation, he thinks they'll provide a "means of reducing our development costs and implementation cycle time by providing more information to developers early in the process." It should also give them a "controlled vocabulary that will greatly help suppliers and implementers communicate using mutually understood and quantifiable terms."
Gas goes wireless
The oil and gas industry is investing more time and money in wireless sensor technologies to help increase production, streamline operations, and reduce expenses, according to a 2005 On World report, "Wireless Sensor Networks for the Oil & Gas Industry" (www.onworld.com). EnCana was one of several big oil and gas companies interviewed. Others included British Petroleum, Chevron Texaco, ExxonMobil, Occidental Petroleum, and Lyondell-Citgo.
Out of 150 companies surveyed, representing oil and gas, OEMs, systems integrators, and component suppliers, 83% use wireless for remote monitoring, primarily for pipelines and wellhead monitoring. The report found 32% were either already using wireless sensor networks or planned to in the next 18 months. All but one of the big oil and gas companies were pilot testing mesh wireless sensor networks.
EnCana focuses primarily on the development of resource plays and the in-situ recovery of oil sands bitumen. According the company's Web site (www.encana.com), resource plays are an "accumulation of hydrocarbons known to exist over a large aerial expanse or thick vertical section, which when compared to a conventional play, typically have a lower geological or commercial development risk and lower average decline rate. Once identified, resource plays have the potential to make a material impact because of their size and low, steady-state decline rates. Applying appropriate technology and program execution are the keys to unlocking value from resource plays."
Initiated in 1999, EnCana started it first resource play in Greater Sierra in northeast B.C., where horizontal wells using under-balanced drilling techniques target the Jean Marie geo-logical formation. EnCana's U.S. operations focus on exploiting deep, tight, long-life, unconventional natural gas formations. EnCana first entered the U.S. by acquiring a stake in the Jonah gas field, which produces an average of nearly 75 million cubic feet per day of natural gas.
Wireless plays in automotive
Patrick Kinney has heard "scuttlebutt over the last year the automotive industry would make a ZigBee play," but he hasn't seen anything definitive yet.
"Now I think we're seeing a lot of focus on ZigBee as a short-range device with low-cost, long battery life for machine-to-machine applications," said the consultant from Kinney Consulting and vice chair of the IEEE802.15 standard. ZigBee is a wireless protocol built on the physical (PHY) layer and medium access control (MAC) layers from IEEE 802.15.4.
As far as wireless inside the vehicle goes, a 3 April Information Week article by Nicolas Mokhoff reported on wireless plays at the Sarnoff Symposium in Princeton, N.J., where researchers at HRL Laboratories and General Motors R&D proposed the build-out of intra-vehicle wireless automotive sensor networks based on the IEEE 802.15.4 standard to make communication between the sensors and control units inside a vehicle easier. "In-vehicle wireless sensor networks would enable the elimination of most, if not all, of the 50 kg of wiring harnesses used in a typical production vehicle," Mokhoff said. "But questions remain about whether the wireless nets would deliver the same performance and reliability as wired subsystems."
Researchers at NEC Laboratories America also looked at using IEEE802.16 (WiMax) for intra-vehicular high data rate communication. "There are two other standards in development for this market," Kinney said. IEEE 802.20 specifies the PHY and MAC of an air interface for interoperable mobile broadband wireless access systems. The IEEE 802.11p task group will define enhancements to 802.11 required to support Intelligent Transportation Systems (ITS) applications. This includes data exchange between high-speed vehicles and between these vehicles and the roadside infrastructure in the licensed ITS band of 5.98 GHz.
SOURCE: "Wireless Automotive Communications," by Thomas Nolte and Hans Hansson of the MRTC, department of computer science and electronics at Mälardalen University in Västeras, Sweden, and Lucia Lo Bello, RETISNET Lab, department of computer engineering and telecommunications at the University of Catania in Catania, Italy. This stems from the proceedings of the 4th International Workshop on Real-Time Networks in conjunction with the 17th Euromicro International Conference on Real-Time Systems (July 2005).
ZigBee, Bluetooth details
ZigBee is a low-cost, low-power wireless public access network (PAN) protocol, intended to meet the needs of sensors and control devices. Typical applications do not require high bandwidth, but do impose severe requirements on latency and energy consumption. The use of such legacy systems raised interoperability problems, which ZigBee technology solves, providing a standardized base set of solutions for sensor and control systems. The ZigBee Alliance (with over 120 company members) ratified the first ZigBee specification for wireless data communications in December 2004.
Bluetooth (IEEE 802.15.1) provides network speeds of up to 3Mbps and was originally devised for PAN deployment for low-cost, low-power, short-range wireless ad hoc inter- connection. The technology has fast become appealing for the automotive environment as a potential automotive wireless networking technology. In 1999, the Hands-Free profile was the first of several application level specifications from the Car Working Group, which the Bluetooth Special Interest Group formed in response to automotive manufacturers' interest.
Using the Hands-Free profile, products that implement the Bluetooth specification can facilitate automatic establishment of a connection between the car's hands-free system (usually part of the audio system) and a mobile phone. Bluetooth wireless products incorporating these enhancements enable a seamless, virtually automatic interface between the car and wireless product. Bluetooth allows hands-free use of a mobile phone either through the car's audio system or wireless headsets, resulting in better sound and control and a safe solution to legislation banning mobile phone use while driving.