Real-Time To The Rescue
While process automation is not a new concept in the food industry, advancements in sensing, monitoring and controlling process parameters continually improve product consistency, process productivity, production flexibility and overall plant efficiency. However, smooth integration between individual systems is just as important as the individual systems themselves.
Mead Johnson Nutritionals' facility in Evansville, IN, is a clear example of this seamless integration. The company recently upgraded its control systems to add flexibility, reduce process variability and identify material loss points in an effort to optimize operations.
The plant produces 15 liquid formulated nutritional food products for two vastly different markets -- infant formulas including Enfamil, ProSobee and LactoFree, and adult nutritional formulas such as Boost, Isocal and Sustacal. Consistency is critical for both markets, as these products may be the sole source of daily nutrition.
Since 1981, a series of process control upgrades were made with off-the-shelf systems and customized communication protocols to tie various systems together. The last upgrade before a total reconfiguration last July was a PC with data acquisition software connected to the main PLC (programmable logic controller). This gave operators rudimentary graphical views of the process but provided no control interface. "Process control was a hodge-podge of integrated systems installed at different times," says Bill Smith, maintenance facilitator of electronics and process controls. "With changing functionality over the years, some devices were added and removed. Drawings and documentation did not keep up with the changes, making troubleshooting difficult, and much of the wiring from the original control systems was never removed."
In July 1998, after nine months of preparation work by Malisko Engineering (St. Louis, MO) and the plant's project team, the plant totally revamped its process control system. Products from various vendors including: Allen-Bradley's (Mayfield Heights, OH) PLCs for base level control; Intellution's (Norwood, MA) FIX and Visual Batch software for control interface and recipe management; Total Control Products' (Melrose Park, IL) remote terminals; and others were integrated together for the total solution.
Mead Johnson's goals were far reaching and demanding, according to Smith. "We required a system that would eliminate barriers to new product adaptability, minimize process variability, allow for future system expansion and give operators real-time access to process parameters, all while assuring production was not done for more than two weeks during installation. The project was based on competitive need versus ROI [return of investment] expectations."
"We have been running on the new system for about six months now and already have seen benefits," notes Smith. "We were able to complete an unplanned new product release in only one and a half days versus the three weeks this type of project would normally take."
Mead Johnson is currently installing its next on-line productivity improvement project. "We are literally in the midst of installing a series of photocells to count cans at various locations on the process and packaging conveyors in an effort to determine and minimize can loss," says Smith. Using an in-house design software system, Smith intends to pinpoint trouble spots, such as filling, sealing, sterilizing, labeling or palletizing. He will also be able to differentiate between loss caused by machine malfunction and other causes.
Moisture-checkDetermining crucial information quickly and continuously, then reacting to that information is a basic benefit of process automation. In bakeries, for example, knowing exact moisture content and temperature of bread or cookie dough prior to entering the oven can allow operators or automatic control systems to positively affect product outcome. On-line monitoring of exiting product parameters can also provide feedback to oven controls and continually optimize oven performance and product outcome.
Traditionally, operators follow a preset formulation and procedure for dough preparation. However, variability within ingredients (e.g. moisture and protein levels in flour, flour and water temperature, etc.), length of "floor time" that a dough sits before depositing, ambient conditions inside and outside a plant and other factors affect optimum oven time and temperature. These factors are critical to gas development within the products, shelf life, color development and other quality parameters.
Alert mixer and oven operators must be constantly attentive, taking physical samples, testing for variables and then adjusting the time/temperature relationship of the oven, hoping to improve or maintain end-product quality. This is not an easy task, even for experienced operators. Austin Quality Foods' Cary, NC, bakery uses an on-line moisture detection system from Dantec (Waterloo, ON) with control systems from Honeywell (Ft. Washington, PA) to relieve operators from some of these tedious tasks.
Austin is the leading sandwich cracker maker in the U.S., producing more than 165 million pounds of sandwich crackers and cookies per year on six production lines at its 400,000 sq.-ft. Cary plant. On its newest line -- a 300-ft.long sandwich cracker production line -- the company examined various methods of improving process control.
"Moisture is the most critical control point in our baking process," says Herbert Killian, senior vice president of manufacturing. "It affects shelf life, color development and other quality attributes. Sandwich crackers are compound products and we are concerned with moisture migrating from the cracker portion into the fat-based filling. We needed a system that could monitor and control cracker moisture to a precise setpoint of 2.5 percent."
The Dantec/Honeywell system is connected to an Allen-Bradley PLC and Panelview graphical user interface. Austin's system consists of an on-line moisture sensor with infrared device for temperature compensation, and model-based computer software.
The 36-inch long, four-inch wide Dantec non-contact moisture sensor operates on the principle of capacitance. Mounted at the oven exit, the sensor emits a radio frequency signal, which is absorbed by moisture in the crackers. Model-based software then calculates moisture based on changes in the electromagnetic wave field. This information is fed to the PLC, which regulates burners, via BTU output, in the direct-fired oven.
"The on-line system has substantially reduced moisture variability in our baked crackers," notes Killian. "While line operators still take periodic moisture samples to verify and calibrate Dantec readings, they seldom need to interfere with the automated control. Line efficiency is 6-7 percent higher than our other production lines due in part to the moisture control system."
20/20 ForesightHow do you improve on real-time feedback and control? Predict the future.
British Sugar, Wissington, UK, is doing just that in several of its nine sugar beet refining processing plants. Using a Connoisseur toolkit from Predictive Control Ltd. (PCL, Northwich, U.K.), the company is saving about $156,000 per year, with an ROI of 1.4 years by anticipating process variability.
"We need to reduce water content from spent sugar beets from about 85 percent water to about 11 percent so it can be sold as animal feed," says Chris Hurd, operations manager with British Sugar. "Our biggest energy cost is in drying the spent product -- costing up to $6.6 million per year for our group of plants." Using horizontal rotary cylinder dryers, British Sugar operated a PID (proportional, integral, derivative control) control system to monitor moisture loss but had to contend with a 15-minute lag time whenever changes were made to drying parameters.
"The predictive control system measures flue gas temperature then models a relationship of the input and output temperatures," notes Hurd. "From these readings, the model predicts future behavior based on current and past behavior, implies final moisture content, then develops a control strategy that determines optimum control settings."
Results include dryer energy cost reduction by 1.2 percent, product yield increase of 0.86 percent, off-spec product reduced from 11 to 4 percent, and downstream processing problems and energy costs significantly reduced. "The biggest benefit is reduced variability," says Hurd. "We no longer have to over-dry product to assure it all reaches moisture specs and our customers no longer have to contend with product moisture variations of two percent which can greatly affect their operations."
Real-Time RumWhether it is sandwich crackers, sugar beets or infant formula, on-line process control yields efficiency and bottom-line improvements to any plant of any size in any industry segment.
Building sales worldwide prompted rum producer, R.L. Seale & Co. of Barbados to do a bit of building of its own. Siting its new facility on the grounds of a 260-yr. old sugar factory dating back to pirate days is only one unique aspect of the plant. It is also one of the most modern and high-tech operations of its kind, combining old world traditions with state-of-the-art real-time process control.
"We run two million liters per year in a fermentation process using a 'closed' tank design," says Richard Seale, CFO. "This controls culture seeding and reaction temperature, optimizing production. With this, our continuous distillation process uses half as many columns to produce as much high-quality product as other comparably-sized facilities while reducing our energy consumption."
The closed tank design also allows carbon dioxide to be recovered, liquefied and sold to carbonated beverage bottlers. In all, Seale estimates that his operation is 30 percent more efficient than similar ones.
Process control is a major point of difference at R.L. Seale. Other distilleries generally choose dedicated controls, where each stage of production is essentially autonomous, according to Seale. "We wanted to have a 'window on the process,' as well as plant-wide integration of instrumentation, simple upgrading and ease of installation," notes Seale. "We went with PC windows-based control that combines features and benefits of PLCs and DCSs [distributed control systems]."
The Freelance 2000 system from Elsag Bailey (Wickliffe, OH) uses an off-the-shelf PC as the operator station, which then interfaces with a process station containing plug-in/plug-out modules. The module configuration determines its monitoring and control capabilities.
"Our system displays mash preparation, fermentation, distillation and spent wash evaporation processes as well as utilities," notes Seale. "The operator station offers a comprehensive process overview with real-time process variable readings for constant monitoring and immediate correction in case of process deviation."
Editor's Note: PCL, operating in Europe, was recently incorporated into Simulation Sciences Inc. (SIMSCI, Brea, CA), a member of the Siebe Group of companies.
Sweet SuccessPermeate Refining Inc., Hopkinton, IA, runs the only facility in the U.S. that converts waste sugars and starches to 200-proof fuel-grade ethanol for gasoline. And it is now running its plant better thanks to a recent process control system installation. The company increased output from 3,000 to 7,000 gallons per day (233 percent) just by automating its distillation process. Further, natural gas consumption per production gallon has been cut in half, and some $50,000 per year in operator mistakes and downtime has been eliminated. Payback time for the $80,000 process control system, from Fisher-Rosemount (Austin, TX), and necessary instrumentation was only three months.
"We scour food and beverage manufacturers for off-spec and contaminated products containing high concentrations of sugar and starch," reports Wayne McCalley, Permeate's process engineering manager. "Recent receipts include millions of peppermint candies; bins of marshmallow; drums of powdered drink mix contaminated in a highway accident; and corn starch from a wet miller. Other popular wastes include corn syrup, fructose and dairy whey. Sugar products are especially desirable as they can be directly fermented to 13 percent ethanol. Complex carbohydrates must first be converted, using enzymes, to simple sugars."
Sticky SituationMimic panels, dial gauges, chart recorders and hand valves initially ran the plant's second-hand distillation equipment, requiring constant operator attention. Even then, the process would swing wildly. One operator ran the process one way, the next operator another. Optimum settings were impossible to determine.
McCalley notes that simply adding transmitters and single loop controllers would have been a half-measure at best. On the other hand, the plant was too small for a conventional DCS, and a PLC/HMI system would have taken too long to develop from scratch. Instead, McCalley installed the new Fisher-Rosemount DeltaV scalable process control system.
Distillation was attacked first. Investigation showed that a 250-point scalable control system plus strategic temperature and pressure transmitters, and motor-operated valves could be purchased for $80,000 -- a small fraction of a conventional DCS. It was, however, a large amount of money for Permeate Refining, and was perceived as a considerable risk.
The Windows NT, plug-and-play, scalable system at Permeate consists of a rail mounted controller and I/O modules, two PC workstations, a redundant Ethernet network tying the components together and graphical configuration software.
Logic Improves ControlSFC (sequential function chart) logic now provides the distillation process with solid interlocking and one-button starts and stops. Every start is exactly the same and proceeds as fast as the process will allow. The logic automatically recycles material already in the still, brings in raw fermented solution (beer) and begins to deliver finished product.
The operator no longer runs around opening and closing valves and trimming temperatures and pressures. Pumps run only when required. The company formerly spent $10,000 per year rebuilding pumps damaged from cavitation or deadheading.
Process Lined Out PerfectlyDistillation currently has 14 loops and a single operating graphic so operators can monitor the process while walking past the control room window. The graphic also provides the engineer with daily and weekly ethanol totals, information the company before could only estimate from storage tank sight glasses.
Real-time BrixEventually, the scalable automation will run -- and tie together -- all plant areas and equipment. It will include raw material mixing, enzyme conversion, fermentation, an ammonia refrigeration/ice storage plant, a boiler plant, reverse osmosis for dead yeast drying and truck filling.
Permeate recently implemented a process link between fermentation and distillation, represented by a Micro Motion (Boulder, CO) Coriolis meter. The meter's primary task is to continuously determine and integrate ?Brix (sugar content) of beer piped to distillation. But it also provides -- either directly or through built-in calculation -- mass flow, totalized flow, temperature, density and proof from a single set of wires.
Freezing FlourWhile most automation discussion focuses toward process line control, maintaining consistent ingredient parameters can be just as critical to optimizing an operation. On the ingredient inflow side of the equation, Bama Pie's Tulsa, OK, frozen dough plant uses a flour chilling system designed by BOC Gases (Murray Hill, NJ) in partnership with Shick Tube Veyor (Kansas City, MO). Controlled by a Honeywell proportional controller (see Food Engineering, October 1998, p. 136), the system precisely monitors and controls flour temperature as it is pneumatically conveyed from external silos to the mixer. Direct CO2 gas injection in the pneumatic line chills flour to 42° F with an accuracy of 0.5° F. As the major ingredient in the mix, flour temperature can significantly affect the final product quality when it is ultimately baked out at the end user.
Bama recirculates the spent air/CO2 vapor into the flour silos, reducing moisture, oxygen and temperature within the storage tanks, helping to inhibit mold growth and achieving a double-win scenario.