Comment

1 November 2006

Say, you want a revolution?

By Rob Conant and Kris Pister

The same pressures that led to 18th century developments are reasserting themselves in the collision of the two worlds of industrial and information technology.

Globalization, rising energy prices, and a strict regulatory environment are driving companies in a wide variety of industries to cut costs while increasing efficiency and productivity.

Changes in organizational infrastructure and processes in their facilities are transpiring in order to stay competitive.

Meanwhile, trying to meet production and profitability goals remains a significant organizational hurdle. Given these market dynamics, industry leaders are feeling pressure to adopt new technologies that will help them gain competitive advantages wherever they can find them.

These forces are contributing to a wave of innovation in how things take place in the industrial sectors we have not seen for many years. The same pressures that led to 18th century developments by inventors such as Eli Whitney, James Watt, and Charles Babbage are reasserting themselves in the collision of the two worlds of industrial and information technology (IT).

Just as the seed press and the steam engine catalyzed industries, so too, are today's manufacturing innovations—particularly those that can provide more access to more knowledge to plant managers and process engineers. This convergence of industry and information is leading to the "New Industrial Revolution."

Imagine, for example, if we could measure, manage, and refine environmental information and other data from the physical world with the same reliability as wired networks but at a lower cost. Data from the physical world, including temperature, lighting, humidity, energy consumption, and movement, could then be married to the world of industrial systems and IT.

Wireless sensor networks and industrial systems are now converging and giving rise to greater efficiencies not experienced since the first Industrial Revolution. As industrial plant managers discover they can do more with less, they are turning to wireless sensor networks that seamlessly integrate with legacy plant systems to develop comprehensive monitoring and control strategies.

Wireless sensor networks allow industry to collect information with more monitoring points, providing awareness into the environmental conditions that affect overall uptime, safety, or compliance in industrial environments and enabling agile and flexible monitoring and control systems. Wireless sensor networks connect critical processes or assets with the systems or experts that can interpret the data or take immediate action. At the end of the day, operational teams with more visibility into their processes can prevent shutdowns and increase efficiencies while reducing the total cost of data acquisition. All of this can add up to a distinct competitive advantage and a head start in the New Industrial Revolution.

Why wireless, why now

The New Industrial Revolution is possible because of recent advances in wireless networking technologies that leverage the capabilities of the existing monitoring and control infrastructure in areas not possible before, more than doubling the available monitoring points at a lower cost per point than current wired solutions.

Wireless sensor networks are comprised of battery-operated motes that have the ability to quickly form a network and communicate with each other. They deploy in a full mesh networking topology where each mote is a router, ensuring extremely low power and achieving greater 99.9% reliability.

Through techniques such as time-synchronized communication and frequency hopping, wireless sensor networks approach the reliability of wired networks, significantly knocking down the barriers to collecting information from the physical world by field intelligent devices.

While the potential to marry physical monitoring with wireless has always existed, the adoption of wireless technology, in particular, has been slow in making its way into industrial-grade monitoring and control systems. Many organizations have discovered traditional point-to-point wireless networks are prone to failure when faced with the challenging and dynamic radio frequency (RF) landscape presented by commercial and industrial environments. Likewise, wireless sensor networks designed for consumer-grade applications such as home automation, PC peripherals, and remote controls are simply inadequate for industrial applications.

However, with the recent technological advances in wireless networks, a whole host of monitoring and control applications—such as oil and gas, cold chain, and machine health monitoring—can now be enabled by seamlessly integrating wireless sensor networks into existing plant infrastructures. The flexibility and adaptability of wireless lowers the physical and cost limitations posed by wired systems, thereby lowering the total cost of ownership of an adaptive control strategy.

Reliability for harsh environments

The measure of success for an industrial-grade wireless sensor network is not how any individual network device performs, but how the system as a whole ensures a reliable flow of critical data. Reliability is an absolute requirement for any monitoring technology because if the data is not reliable then the economic benefits of its low installation costs are irrelevant.

Specifically, for a wireless technology to be reliable in industrial applications, it must function in harsh industrial environments with unpredictable electromagnetic interference, radio frequency (RF) fading, and multi-path interference. It must also co-exist in the field with other wireless devices or noise emitters such as machine equipment, communications devices, walkie-talkies, instant connect phones, pagers, cell phones, remote controls, and other wireless frequency emitters common in the industrial environment.

New wireless sensor networking topologies are one factor that drives new levels of reliability. In a mesh network topology, each mote has at least two parent motes with which it can communicate. Even if an individual link becomes inoperable, a mote still has a communication path available. This redundant routing ensures resiliency in case of offline motes or broken links.

Wireless sensor networks using the Time Synchronized Mesh Protocol combine frequency hopping spread spectrum (FHSS) transmission and time synchronization, varying communications in both frequency and in time to sidestep RF interference problems. This technique ensures alternate paths are available if RF interference blocks any signal. FHSS technology is particularly useful in industrial environments where intermittent RF interference is common.

Wireless sensor networks provide adaptive monitoring systems in industrial environments with the flexibility needed in a plant's monitoring and control strategy. Wireless motes can go where needed without the need of specialized RF skills or site surveys, while the network handles the rest, such as wireless connectivity, routing redundancy, and frequency agility. Additionally, wireless sensor networks are adaptable to changes in both the configuration of equipment on the plant floor and in the layout of the network itself. If managers add or remove monitoring devices, the network simply reconfigures itself automatically.

As wireless networking technology advances, it is also becoming more cost-effective. Current wireless sensor network designs ease development and integration with other systems. No customization, integration, or development is required, and there are no wiring or installation costs. Battery-powered motes do not require AC power, which can make wireless networks suitable for locations where power distribution for additional monitoring equipment does not exist. Because motes self-organize into a functioning mesh network, no site survey or wireless expertise is required, and the installer does not have to program or configure the devices.

Wireless sensor networks also deliver long lives with a minimum of ongoing maintenance. Motes can have a lifetime of five to seven years on a single pair of AA batteries depending on the application, and they report on their power status, so operators or managers can tell when battery replacements are due.

Impacting prime applications

The value of wireless networks is becoming apparent to organizations who have found they need real-time access to knowledge about their plant's environment, processes, and equipment in order to prevent disruption. Spurred by the recent technological advances, the New Industrial Revolu-tion is beginning to influence three very distinct monitoring and control applications: oil and gas, cold chain, and machine health monitoring.

Oil and gas manufacturing is one industry looking toward wireless sensor networks to solve critical information gaps. Large industrial sites, such as oil refineries already have complex process-control systems in place, but there are many additional points that could provide additional data to optimize processes. However, because of the complexity and cost of integrating these non mission-critical points into the existing control architectures, many of them are not. Monitoring is usually a manual task with inspectors checking the status of key motors, valves, pumps, and supporting process variables. Not only is this time-consuming, but when multiplied across an enterprise's global infrastructure, it also becomes very expensive. Dedicated staff inspecting the refinery may be able to detect a problem, but most likely, only after the problem has occurred.

Wireless sensor networks reduce the oil and gas industry's reliance on inspection personnel by supplying information about utilization rates, energy usage, equipment conditions, and environmental conditions. Infor-mation derived from wireless networks can help quickly identify and address issues that could jeopardize overall uptime, safety, compliance, and profits. The network drastically reduces the cost of accessing information per point, and the ability to gather information around the clock provides unprecedented monitoring capabilities.

As well, companies in the food, chemical, and pharmaceutical industries have unique business problems and monitoring requirements. The shrinkage of perishable product inventory can range from between five and 10%. In real dollar terms, that could translate into hundreds of millions of dollars in losses at a large grocer, for example.

Companies in these industries depend on the cold chain—the monitoring of environmental conditions such as temperature and humidity throughout the supply chain—in order to prevent such losses. These conditions have a major effect on freshness, and poor cold chain management results in spoilage and other conditions that make products unsuitable for stocking.

Financial considerations notwithstanding, the cold chain is also crucial to the food, chemical, and pharmaceutical industries' compliance efforts. For example, federal penalties can be thousands of dollars per instance when regulators find perishables have not been stored within federally mandated levels to ensure food safety.

The risk of drugs stored in inappropriate environments can be orders of magnitude higher. Wireless sensor networks can provide actionable information about environmental conditions in physically challenging environments such as a cold storage container, a large distribution facility, or in a hospital.

Information on temperature, hu-midity, and carbon dioxide levels alerts managers that preventative action is necessary. It can also provide evidence of compliance to regulatory bodies, thus avoiding fines, and alerting managers to equipment that may be outdated or operating at less than optimal levels.

The resulting savings are critical, and the low-cost nature of wireless sensor networking eases adoption barriers.

In many industries, and especially manufacturing, equipment performance is of critical importance. The maintenance of machine condition or operation has traditionally been a manual enterprise. Bearing problems are one of the most common faults in industrial machines, causing unplanned downtime on essential production equipment.

Typically, monitoring vibration to detect bearing problems involves manual recordings or requires expensive systems or services. Another aspect of machine health monitoring is temperature monitoring to detect abnormal or sub-optimal machine behavior in order to prevent equipment from further deterioration.

Wireless sensor networks continuously collect information about the condition and behavior of machines within plants. Those responsible for equipment benefit from access to expanded monitoring information at greatly reduced costs, allowing them to prevent disruptions, achieve greater uptime results, and meet production goals.

So, this revolution is the long awaited convergence between industrial systems and reliable, flexible wireless sensor networks.

Today's plant managers and engineers can now expand the capabilities of existing monitoring and control solutions to new applications and be free from the physical and cost limitations of wiring to achieve affective adaptive control strategies.

The ramifications for the industry as a whole are clear: greater insight into machine behavior, less unplanned down time, improved data analysis, and a safer and more productive working environment, which translates into dramatic improvements in plant operations, reduced costs, and increased competitive advantage.

ABOUT THE AUTHORS

Rob Conant is the co-founder and a vice president at Dust Networks. Kris Pister is a professor of electrical and computer engineering at the University of California, Berkeley.