Not Your Mama's Sensor

More rugged, adaptable and reliable sensors are being deployed in industrial automation. The fact they’re taken for granted attests to how well they perform.

When the Lexus LS 460 debuted, its self-parking system created much of the initial buzz. Formally known as the Advanced Parking Guidance System, the sensor-driven device automatically eases into parallel-parking spaces. 

Of course, it didn’t take long for the automotive crowd to laud the programming and discount the sensors as the same technology that’s been around for decades. After all, sensors are the Rodney Dangerfield of the automation world: electronic systems don’t work without them, but they still get no respect. If proximity and photoelectric sensors are so important, critics sniff, why doesn’t anyone notice them?

In fact, being taken for granted may be the greatest complement for today’s technology. Like other electronics, sensors are more robust, easier to use and less expensive than ever before, and food companies and other manufacturers are taking advantage of those improvements to develop new applications that reduce waste and improve processes. Faster, smaller, lighter and cheaper is the mantra of sensor fabricators, observes Jim Pankiewicz, sensor marketing manager at Schaumburg, IL-based Omron Electronics LLC, and advancements in all four areas are opening up new possibilities in inspection and quality assurance throughout the supply chain.

Sensor technologies that may have been reserved for medical and other advanced applications are finding their way onto the plant floor. Ultrasonic sensors are a case in point. Flat beer was the top complaint at one major brewery, recalls Carl Bonnan, general manager at Heuft USA Inc., the Downers Grove, IL, division of a German firm specializing in automated quality-assurance technology. Internal pressure is a reliable indicator of lost carbonation, but non-destructive measurement posed a challenge. Engineers used an ultrasonic sensor to measure the magnetic field above the crown, using amplitude, frequency and another undisclosed parameter above the crown to extrapolate pressure. “That sensor has pretty much eliminated flat beer as a consumer complaint,” maintains Bonnan.

The logarithm that calculates pressure based on the sensor’s inputs may be the genius of the system, but the sensor is the enabling technology.  “Years ago, manufacturers just did compliance inspection,” Bonnan observes. “Now people want to know how good or bad the product is so they can zero in on an issue before it becomes a problem.” The savings from reduced waste are enormous, and sensors invariably underlie the solution.

Sensors are stand-ins for human senses, and no where is this more apparent than with photoelectric devices. A low cost option to complex vision systems, photoelectric sensors are penetrating new areas of the plant and expanding the ability to manage processes. One consequence is the introduction of beefier, detergent- and moisture-resistant sensors. Omron introduced a 316 stainless steel housing for its E3ZM photoelectric sensor last year, an upgrade driven “almost exclusively by food & beverage because of the stringent cleaning regiment,” says Pankiewicz. “We try to mimic what our eye tells us with these sensors: is it color, distance, or orientation we’re trying to sense?” Photoelectric devices can handle all these tasks, but they also must withstand rapid temperature cycling and other industrial abuses.

Like Omron, Rockwell Automation is seeking IP69K certification for heavy-duty sensors, according to J.J. Thiara, marketing manager-photoelectric sensors. “It started out with proximity/ inductive sensors and is expanding to other units,” he says.

Photoelectric sensing range has gradually increased, “but a lot of times it isn’t the range but the precision that a customer wants,” Thiara says. As background color and contrast changes, sensors can be fooled; unlike the human eye, they can’t adjust to changing conditions. Nonetheless, flexible manufacturing demands adaptability, Thiara notes, and simplified teach functions are being engineered into some sensors to adapt the focal plain as needed.

A teach function that can be executed by a low-skill operator or remotely is a growing demand, Pankiewicz agrees. On-board ASIC (application specific integrated circuits) components are compact and affordable enough for inclusion in “mildly complex sensors,” he says. These microprocessors also can store multiple settings to adapt to line changeovers.

Multi-function sensors are another notable development. Instead of requiring two or more inputs to a PLC, these units can perform logic, timing, counting and other functions with a single device. “By reducing the I/Os on the PLC, you also reduce the lines of code that have to be written,” Pankiewicz points out.

Look, ma, no wires

Wireless sensing is the new frontier in industrial sensing, with reliability the caveat most often raised. Some believe reliability no longer is an impediment: power source is the bigger hurdle, particularly for active devices.

Nine years ago, Canada’s Sensor Wireless Inc., Charlottetown, PEI, introduced SmartSpud, the first in a series of wireless sensors that measure impact on a vegetable or packaged food product during processing. Those early devices relied on a C battery to transmit data to a device 20 ft. away, recalls Tom McDonald, senior product development technologist. Today, a cell similar to a digital camera’s battery powers the sensor and transmits results to a personal digital assistant up to 50 ft. away.

Four years ago, the firm’s technicians created their first sensor for monitoring impact levels on a bottling line. An exact replica of a client’s bottle was created and hollowed out, with sensors placed at the shoulder, heel and other impact points. Instead of storing data for later download, as a data logger would do, the device used radio frequency to provide real-time readings as the mock bottle made its way down the line. “By capturing acceleration impact at five points, bottlers can make preemptive changes to a line each time they lubricate or make some other change, instead of reacting to maintenance issues,” McDonald says.

Use of the ersatz bottles is taking off in the UK, thanks to a lightweighting campaign funded by the government and encouraged by major retailers, notably Tesco. The goal is to slash 65,000 tons of glass and 48,000 tons of related carbon emissions from the supply chain within two years by reducing the thickness of glass containers. Glass manufacturers specify the impact standard of their lightweighted bottles, Sensor Wireless technicians correlate the standard to G-force, and the mock bottles record what actually transpires when the bottles roll down a line.

Within 50 ft. of the replica bottle, 99.6%-99.7% of transmitted data is captured, says McDonald. “Nobody can guarantee 100% of the data, and that’s why there’s a lag in getting QA to convert from pen-and-paper data logging.” For process control, the gap is negligible and easily corrected by simply running the replica through the line again.

The next goal for Sensor Wireless is to incorporate GPS with their replica containers to relay data as product moves through the distribution system. They will be following in the path of firms such as Novazone Inc., a Livermore, CA, provider of ozone-based applications for food companies. Novazone recently launched PurFresh, a sensor-based system that adds measured amounts of ozone to a refrigerated truck or shipping container to slow ripening and kill mold and bacteria during produce transport. Sensors measure temperature, humidity, ozone levels and other factors and relay the data via LAN to a control unit built into the cover of the refrigeration unit, according to Michael Weber, Novazone’s vice president-engineering.

“Reliability of the sensor network was a concern; there are lots of home networks available, but in an industrial environment, even cell phone reception or BlueTooth can be spotty,” says Weber. Novazone selected a time synchronized mesh protocol from Dust Networks to provide redundant data networking (see sidebar on page 87). “In a refrigerated container, power is not an issue,” Weber adds, making reliability the top design consideration.  

Nanoscale sensing

Biosensors have had an up and down relationship in food, mostly down. A recent example is Advanced Biosensors Inc., a Hunt Valley, MD, venture that introduced engineered biosensors that emitted light from a calcium sensitive, bioluminescent molecule when a specific pathogen attached itself to the biosensor (“A better germ-detecting mousetrap,” Food Engineering, May 2005). B-lympocytes to detect E. coli, Lysteria and other organisms were developed to deliver reliable results and few false positives in a matter of hours, not days as with conventional biodetection cultures.

Advanced Biosensors targeted the food industry when its biosensor technology was rolling out commercially two years ago, but recently the company has shifted to anthrax detection and other biohazards. “There doesn’t seem to be much impetus to do wholesale food testing,” says Rick Thomas, vice president-business development. Advanced Biosensors recently “made the strategic decision to move away from the food area and move toward defense and biotech,” he says.

An amino assay of a different sort has been under development for almost 20 years at the Georgia Tech Research Institute (GTRI). Its developers remain firmly committed to the food industry, in particular the poultry segment. GTRI’s biosensors employ optical waveguide interferometers to detect the presence of viruses and pathogens. Rapid detection with a low-cost test always were twin objectives, and a year and a half ago, researchers upped the ante and began developing an in-line system, according to David S. Gottfried, the senior research scientist heading the project. That would transform the biosensor from a lab-based instrument to an operations control tool, he says.

A reusable optical chip needs to be developed to make the biosensor commercially practical. The rest of the electronic hardware is readily available for less than $1,000, and detection levels below 500 cells per milliliter have been demonstrated. Field tests for Avian Influenza (AI) detection and chill-water quality may be done this summer. AI testing with deliberately infected chickens would have to be conducted in a Level 3 biohazard chamber, and that is problematic. Pathogen detection in chiller water “is probably easier to field test,” Gottfried acknowledges, bringing that application of the technology closer to reality.

Processors typically sample chiller water weekly because testing is costly, labor intensive and time consuming, he says. By providing an inexpensive way to monitor bacteria levels and giving plant operators a clearer idea of what is going on in the chillers, researchers hope to enhance chiller management systems that use less water, minimize the likelihood of foodborne outbreaks and reduce the cost of excessive use of disinfectants.

“I’m a chemist, and while we’ve resolved some of the chemistry issues, there are a number of engineering issues that still need to be resolved,” admits Gottfried. “The research is still somewhat in its infancy.” Still, development of a working model is progressing, and refinements and optimization will continue long after the first commercial biosensors become available.

The same kind of continuous improvement applies to wireless, photoelectric and other types of sensors. The culmination of little improvements add up to significant advances. Lexus’ APGS seems like whiz-bang technology now, but it requires the driver to first align a yellow flag on a screen with a box representing a parking space at least 78 inches longer than the car before easing up on the brake and letting APGS go through its paces. The day when drivers can simply exit and car and bark, “Park!” arrives, people will applaud the on-board computer and say, “The sensors are the same as they were in 2007.”

For more information:
Rob Conant, Dust Networks, 510-400-2900

David S. Gottfried, Georgia Tech Research Institute, 404-407-8300, david.gottfried@gtri.gatech.edu

Carl Bonnan, Heuft USA Inc., 630-968-9011

Michael Weber, Novazone Inc., 925-454-0303

Jim Pankiewicz, Omron Electronics LLC, 847-285-7374

J.J. Thiara, Rockwell Automation, 978-441-9500, jsthiara@ra.rockwell.com

Tom McDonald, Sensor Wireless Inc., 907-626-3952, tech@sensorwireless.com

ROI on wireless sensor networks spotlighted

Reliability and power consumption are two of the biggest stumbling blocks for wireless sensor networks. Not only have those issues been resolved, insists Rob Conant, co-founder of Hayward, CA-based Dust Networks, but low-cost solutions are beginning to be deployed in food operations.

“Reliability had been an issue for a long time, and it’s a deeply held belief that it remains to be an issue with wireless,” observes Conant. “With time synchronized mesh protocol, even if one link for data transmission is blocked, wireless networks have four or five other paths through which information can be routed.”

Sensors and actuators for valves and other devices remain tied to a local power source, but sensors for communications networks have solved the power riddle. Mesh networking technology “exceeds 99.9% reliability, even in metal canyons,” he says. “Our engineers’ relentless focus on low energy demand has allowed us to develop products that can go years without changing batteries.” Wireless slashes installation costs and eliminates the “huge rat’s nest of wires if you’re deploying thousands of vibration sensors,” he adds.

Petrochemical companies are converting to wireless networks en mass, and food companies soon will follow: Cargill was sufficiently impressed with mesh networks to take an equity position in Dust Networks.  “They see a ton of opportunities for wireless to improve their operations,” says Conant.