After operating in the shadows backstage for decades, process analyzers, detectors, spectrometers, chromatographs and other instruments are finally getting some recognition in the limelight—mainly because their data is spreading further to be used more effectively.
Quality end products have always relied on maintaining raw material characteristics and specifications throughout their manufacturing process, and this means thorough instrumentation and analysis. However, on-board intelligence, in-the-field data processing, enhanced networking, Internet links and even virtual computing are helping instruments provide better, faster and more widely distributed process analysis, optimize their processes, and deliver better quality than ever. The trick is knowing where and when to deploy these instruments and their increasingly digitized, online and accessible data.
For instance, radiography with traditional wet film has been the primary non-destructive testing (NDT) method for corrosion monitoring and wall-thickness measuring at YPF's Mendoza plant, but the state-run refinery in Argentina recently needed to simultaneously increase inspections and reduce downtime required for each one. As a result, Mendoza's refinery inspection manager, Martin Rebollo, decided to replace the plant's aging film equipment with digital radiography technology, so its onsite inspection team now uses GE's CRXFlex computed-radiography scanner with phosphor plates and an IP-adaptor that allows smaller plates to be scanned simultaneously in a 14 x 17-in. cassette. They also use GE's Flash filters for sharper images and Rhythm software for rapid review and sharing of inspection results, and added a DXR 250C-W portable wireless detector to their digital portfolio (Figure 1).
“We have the same setup as conventional film, but digital is much easier," says Rebollo. "Our inspectors love the fact that there's no need for reshoots, which lets them be more productive.”
In short, switching to computed radiography means shorter exposure times, more inspections per shift and less downtime at the Mendoza refinery. It also means fewer retakes because CRXFlex's wide dynamic range and high signal-to-noise ratio allows a broad range of thicknesses to be inspected in one exposure. And, thanks to increased inspection productivity and savings in consumables, Rebollo reports that the plant's CRCFlex scanner and phosphor plates are scheduled to pay for themselves in two year, or in just six months if the cost of film compared to the phosphor plates is added.
Joe Downey, product manager for Foxboro liquid analytical products at Schneider Electric, agrees that pH, conductivity, dissolved oxygen and other analyzers are mature technologies, and that there's not much new in their fundamental sensing functions, but there's a lot going in the intelligence and communications capabilities being added to them. "Smart sensors and analyzers can carry their own calibration parameters, so many users don't need to add calibrations," says Downey. "They can also do diagnostics on themselves and constantly measure their internal performance, which means users can do better predictive maintenance, and even asset-manage groups of sensors and instruments like a fleet. We're also seeing more multi-channel analyzers that can accommodate four to eight sensors, measure up to 16 parameters in the same pipe or in different locations, and distribute data via digital communications, especially Ethernet over Internet protocol (IP) and wireless. This is getting us closer to a true universal analyzer that can measure pH, conductivity or mix and match other values in one device.
"Some smart instruments are even using onboard intelligence to plug basic measurements into algorithms and secure real-time percentages of acid, soluble oil, water content and other results that they previously could only do in a lab. Next, they're combining sensing and communications to deliver digital signals right from the sensor, and take the usual transmitter out of the loop."
Analyzer Trends Analysis
Despite the online capabilities, IP connections and other innovations added to process instruments and analyzers in recent years, potential users can still be slow to recognize, adopt and benefit from them.
"The refining, petrochemical and natural gas markets are very conservative relative to new technologies, and onerous safety-area design constraints and third-party testing are barriers to entry. For example, an accepted technology such as a zirconium oxide cell to measure excess oxygen in furnace flue gas is still widely used more than 40 years after its introduction," explains Mike Fuller, vice president of Ametek Inc.'s Process and Analytical Instruments division. "Growth in the analytical technologies market in recent years has been about 8 to 10% per year with most growth coming in the BRIC countries of Brazil, Russia, India and China. However, the 50% drop in the oil prices over the last six months has greatly reduced upstream capital expenditure spending with a lesser effect on downstream capital purchases. In the U.S., production of shale gas and liquids has resulted in many new petrochemical plant projects for producing petrochemicals such as ethylene and methanol."
Fuller reports that process customers are favoring low-cost, spectroscopy-based technologies over process gas chromatography that is costly to maintain. He adds that modern analyzers also have digital communications capabilities, including an Ethernet-connected, web-browser-enabled remote interfaces.
"Tunable diode laser absorption spectroscopy (TDLAS) is the fastest-growing technology in process analysis. Ametek secured more than $5 million in orders for our TDLAS analyzers/integration as applied to meeting the U.S. EPA's Sub Part J(a) Flare Gas regulations that are going into effect in 2015. Our analyzer covers a 0 to 60% H2S range and a 0 to 300 ppm range in one analyzer," says Fuller. "The long lifetime of the semiconductor source and the non-contact nature of the approach yield a low cost of ownership compared to other methods. Another area with significant advances is predictive maintenance and other advanced diagnostics. Predictive maintenance algorithms monitor the condition of consumable components, providing future estimates of needed service. This allows analyzer maintenance to be scheduled when the process is in a routine shutdown."
In the future, Fuller projects that widely tunable, mid-infrared, quantum-cascade laser (QCL), semiconductor-based process analyzers are being commercialized and will likely replace many process FT-IR and process gas chromatographs. These devices can quickly measure three to five compounds with just one laser. "Current prices for QCL lasers limit the cost competitive applications today, but prices for the lasers are likely to drop dramatically over the next five years," adds Fuller.
Ulrich Gokeler, strategic support director for process industries and drives in Siemens' Analytical Process Group, adds that automation and intelligence are rapidly proliferating in many sampling systems and instruments. "Analyzers are getting microprocessors and becoming as automated as their plants," says Gokeler. "So maintenance guys and analyzer engineers that know their crucial parameters can benchmark and establish them onsite, program them into their analyzers, and use software such as Siemens Analyzer Management to store and display performance and quality results over time."
Instruments Widen Influence
Of course, more smarts and better communications mean users can rely on their analyzers to help solve more difficult problems.
For example, to minimize formation of trihalomethanes (THMs) in its water and comply with the Stage 2 Disinfectants and Disinfection Byproducts Rule (DBPR), the municipal water treatment plant in Benicia, California, has deployed an online analyzer to monitor hard-to-characterize fluctuations in THM levels throughout the day. The plant is rated at 12 million gallons per day (MGD), but average daily flow is about 6 MGD. Continual monitoring helps improve coagulation and disinfection of residual total organic carbon (TOC) from natural organic matter (NOM), which combines with chlorine gas disinfectant to create THMs, according to Scott Rovanpera, superintendent at Benicia's water treatment plant.
Following earlier filtering, coagulation, source water reallocation, alum sludge removal and disinfectant control efforts, the Benicia plant tested an automated THM-100 online, self-calibrating analyzer with 4-20 mA networking and USB connections for data acquisition from Aqua Metrology Systems. Every four hours, it employs purge-and-trap sampling, desorption into a chemical mixture to generate a colored product, and time-resolved spectrophotometric analysis for detection and determination of THM levels.
"After installing the online THM analyzer and establishing baseline THM data in October 2012, we began to undertake weekly process changes to determine the effect on THM levels," says Rovenpera. "The controlled experiments only included testing of conditional changes that would improve water quality and maintain THM levels consistently under 40 ppb, and the precision and bias of the online THM analyzer yielded better results than traditional laboratory methods and provided operations staff with real-time feedback on the implication of process changes on THM levels in their finished water.
"Since installing the online THM analyzer, our staff has accurately monitored the correlation between operating conditions and THM formation potential; operators can optimize plant performance and chemical dosage, allowing them to economize plant chemicals and dose pace accordingly; and treatment anomalies have been detected early, enabling immediate correction to process controls."
Likewise, Dow Chemical Co.'s Rohm & Haas division in Deer Park, Texas, needed to troubleshoot two similar chemical units showing widely different signs of corrosion damage. While one plant had low corrosion rates, the other corroded at very high rates, causing rapid failure of stainless-steel piping, and traditional monitoring couldn't detect the root cause. Initially, Rohm & Haas' engineers planned a $500,000 alloy upgrade, but then decided to try SmartCET corrosion transmitters from Honeywell Process Solutions and connect them via HART protocol to the units' Honeywell-based DCSs (Figure 2).
The new probes and transmitters gave operators access to real-time, online corrosion data, including time-trended general/uniform rates and localized/pitting corrosion information. They also allowed the corrosion data to be alarmed, historized, trended, assigned and seamlessly correlated with other process variables, providing a broader view of plant operating conditions and methods of mitigation. The SmartCET solution enabled Rohm & Haas' operators and engineers to identify two process scenarios contributing to the difference in corrosion rates.
First, data from the transmitters showed that one unit's corrosion rate was higher immediately after shutdown. Further inspection showed that a leaky valve was allowing water into the system on shutdown and increasing system corrosivity. The leaky valve was replaced, eliminating this source of corrosion. Second, corrosion was higher in the unit's process when a certain recirculation condition occurred. So by making some process changes, the plant avoided this condition on the second unit. When both conditions were corrected, the team learned they could avoid the planned $500,000 materials replacement on the units.
“Having a corrosion sensor that connects to the DCS is a real benefit,” says Andrew Wheeler, Deer Park operations team member at Rohm & Haas. “We can view corrosion just like any other process reading, and this allows us to make decisions that preserve our equipment while producing world-class products.”
CO Sampling for Faster Safety
Unsurprisingly, as process analyzers show how their improved capabilities and communications can optimize production and product quality, other users and supportive developers are using them to meet their requirements.
For instance, Huayang Group's 4,200-megawatt, seven-unit, coal-fired HouShi power plant in Fujian province, China, decided in 2012 to add carbon monoxide (CO) monitors to its five coal mills in Unit 1 to supplement its fire and temperature sensors. Similarly, Hoosier Energy's 1,070-MW, two-unit, coal-fired Merom generating station in Sullivan County, Indiana, burns 10,000 tons of coal per day; runs three double-ended, ball-tube mills per unit; and needed 12 sample points per unit or six twin-stream CO monitors, plus an in situ oxygen analyzer. They were installed on the units' classifiers during a May 2011 outage (Figure 3).
Both power plants selected AmetekLand's Millwatch analyzers, which have rugged sample probes with automatic blowback to maintain good sample flow, automatic calibration to verify correct operation and correct CO response, and continuous measurement of each sample point with no multiplexing and sub-60-second response time. This feature is important because hazardous conditions can develop in coal mills in just a few minutes, while multiplexed systems sampling six measurement points typically only sample each point once every 10 to 15 minutes. For example, the HouShi plant reports that Millwatch detected rising CO levels in a coal mill outlet in 2013, which allowed corrective action to be taken much sooner than the 15 minutes it would have taken the temperature and fire systems to respond.
Likewise, Merom Unit 2 also tripped in 2013, and operators observed a rapid CO increase, even though there was indication of a temperature increase. Within a few minutes, its CO level was above the alarm threshold, and operators decided to activate the unit's deluge system. The boiler continued to operate, using coal from the remaining mills with output dropping to 60% of its rated value. Meanwhile, the CO level in the affected mill started dropping after 15 minutes, and within 45 minutes it was below 10 ppm. The mill was restarted two hours after the high CO alarm was detected, and was returned to full operation in 3.5 hours. Again, Millwatch quickly detected a potentially dangerous coal-oxidation condition and allowed it to be dealt with before a fire or explosion could happen. In fact, neither HouShi or Merom have had explosion since Millwatch was installed more than three years ago.
"Regulated communities have to follow that the EPA and its worldwide counterparts are doing, such as setting rules for monitoring mercury and air toxics (MATs) emissions from power plants, which will be finalized in 2016," says Derek Stuart, product manager for combustion and environmental analyzers at Ametek. "We don't make a mercury analyzer, but existing models use UV spectroscopy. However, mercury is difficult to sample, and these analyzers are hard to calibrate, so developers are looking at different approaches and materials.
"In addition, Millwatch and Silowatch co-analyzers have been available for about 10 years, but we've been able to increase their sensitivity during the past three years. There's also more interest in them because they do continuous sampling instead sequential sampling. Ametek is also continuing to investigate opacity monitors for smoke and in-stack, laser-scattering devices."