Printer Friendly

PC-based SPC.

How can you tell if you're painting an accurate picture of your commitment to quality? You can fool some of the people some of the time, but you'll eventually paint yourself into a corner.

Let's be realistic. If all the companies that claimed to be practicing Total Quality Management really were, the economy would be booming. For some, quality is an ornamental patina applied by management. For others, it's a buzzword used to instill a false sense of security. One thing is certain: Management cannot lie to itself for long before the consumer market tells the real story. While quality is almost by definition an art, the applied science of quality management, called Statistical Process Control, provides the toolset for TQM and other quality improvement paradigms. SPC can be defined as a set of statistical problem-solving tools for identifying, analyzing and eliminating problems. The tools include flow charts, Pareto diagrams, cause and effect fish bone" diagrams, histograms, scatter plots, run charts and control charts. (See accompanying story on page 31.)

SPC one-ups traditional quality control, which focuses on final product inspection and removal of defects for rework or scrap. Total Quality practitioners estimate that 20% to 3 5 % of the value of goods is lost to reworks, scrap and redundant inspection steps. Beyond the old realm of defect rejection lies the current domain of SPC, where data are used not only to re-engineer the product, but to improve the process.

Packaging line SPC

The most common point of SPC application is the packaging line, where volume and weight control are the primary concern. This is due to the potential for rapid return on investment in yield and cost savings. Companies that can reduce the amount of product overfilling and still meet label weight declarations are in a prime position to reap monetary benefits through SPC. Whether a processor is looking for cost savings or product quality, SPC imparts uniformity to formerly untamed operations. This goes for ingredient batching weighing and the application of pizza toppings as well as packaging.

On its Celeste brand pizza production line in Jackson, Tenn., Quaker Oats reduced cheese topping usage by 500,000 lbs per year for a savings of $600,000. The application involved teamwork, computerized controls and on-line data entry into an SPC software application.

One packaging consideration that warrants statistical analysis is closure integrity. If liquid product splashes up on the seal area of films, heat-sealing may be incomplete. And if bottlecap applicator parts wear, age, become contaminated with lubricants or otherwise lose their ability to grip, they can cause difficult-to-detect problems related to removal torque.

Further down the packaging line, SPC has been applied to corrugated shipping container issues, from the shape and die-cut of the box for optimum pallet stacking patterns, to analyzing warpage due to temperature and humidity variations, to the proper gluing and handling of corrugated by case-erecting machinery.

One food processor was experiencing line jams on the automatic caseerecting equipment that packs finished goods into the corrugated shipping container. Variances in corrugated thickness as low as one-tenth of an inch were causing problems. Too thick and the box wouldn't fully enter the case erector's in-feed. Too thin and the carton would travel too far into the machine. Either way, line jams resulted.

It was determined that the packaging supplier's machinery was near state-of-the-art, so switching suppliers was ruled out as a solution. The problem was solved by building greater tolerance into a spring in the food processor's equipment.

"Obviously, the economics were in favor of changing a little spring on the packaging line, as opposed to changing a paper industry standard," explained quality consultant Hans Bajaria, Ph.D., who applied SPC at both processor and supplier sites before the problem was identified.

Root cause analysis

Without identifying the causes of a variance, the SPC practitioner is prone to making continual mechanical adjustments, a practice called "tinkering," that can overwork the system and shorten equipment service life. Bajaria compares tinkering to another familiar experience. "It's like driving a car. If you're always turning and steering to stay in your lane, you'll eventually wear out your tires or destroy parts of the car. It's not enough to correct the problem. You have to ask why your alignment-or process-is drifting in the first place."

"You have to work backwards in the system to find the causes of downtime until you ultimately eliminate the cause," concurs quality expert John Surak, Ph.D., who is affiliated with Clemson University's food science department. "Too often we deal with the symptoms. What SPC does is provide a systematic problem solving tool to find the root cause of a problem." Surak reports having seen companies use SPC tools such as root cause analysis to reduce downtime "from about 90 minutes a shift down to three minutes."

On root cause analysis, T.J. Lipton's Lynn Hare, Ph.D., adds, "SPC will help you on the packaging line, and by all means it should be used there, but it's also important to recognize that by the time you have experienced product variation on the packaging line, it's almost too late. What you really need to do is work to drive quality upstream. That doesn't kill the opportunity for SPC on the packaging line, it makes it easier. What you're doing is providing the packaging line with more uniform product to be filled, and that's going to guarantee or build in the kind of control that you're looking for."

Whether driving quality upstairs or upstream, top administrators need to realize that there is plenty of room for process improvement, and streamlining can be done without slashing budgets.

Campbell Soup practices

supply side SPC

Campbell Soup Co.'s Robert Bottel, director of the company's Select Supplier program, has been part of the effort to improve quality and profitability at Campbell. He says, "Today's Total Quality Management thinking consists of preventive control of processing factors applied as far upstream in the supply process as possible. It demands the use of statistical process controls that insure the goods or services meet specification or expectations as they are being made or performed."

According to Bottel, measurable Select Supplier benefits were $2.3 million 1990, $18.4 million in 1991, and more than $25 million in 1992. "The cost to administer the program has been approximately one-twentieth .. of the total benefits."

SPC by both Campbell Soup and its suppliers has identified problems, targeted solutions and measured gains. By forcing SPC up through the channels of supply, Campbell, its suppliers and its consumers win.

Specific SPC generated gains at Campbell include:

* Savings of $248,000 per year in carrying charges for corrugated cases. Order-filling lead time dropped from 10 to two days as supplier inventory dropped from 4.9 million cases to 1.8 million.

* Reduced variation in the weight of hash brown potato patties; 84,434 extra" patties were unnecessary for a 1 % reduction in cost. A

70% reduction in consumer complaints for defects in frozen, diced meats for soups.

* Savings at a laboratory supplier's site of $180,000 per year due to inventory management innovations.

* Savings of $1 3 5,000 per year for a Campbell supplier of computer and copy paper. Supplier savings are at least partially passed along to Campbell and its customers.

Bottel is planning to present these data to the Philadelphia Section of the American Society for Quality Control on Nov. 18.

Integration of SPC

While after-the-fact data are part of the process of measuring management, SPC also relies on hard technology. Campbell uses on-line laserscanning meters to ensure label tolerances of 1/64" at its Napoleon, Ohio, soup site. Other processors use a combination of on- and off-line observation and electronic data collection to provide a complete picture of the problem-solving environment.

Not only does TQM depend on SPC, but SPC needs TQM teamwork, training and follow-through to succeed. Companies lacking these qualities risk buying into SPC as a fad, working with a computer software application for a couple of months, then watching years pass as the user manuals fossilize on the shelf.

"Implementation of SPC can show the benefit of a modem Quality Management program, but it has to be integrated with Total Quality Management. If SPC is not incorporated within the context of TQM, you may have short-term gains or short-term success in terms of measurable savings, but in the long term it will fail," according to Clemson's Surak.

Lipton's common SPC sense

The $1.5-billion-a-year T.J. Lipton company has faced both critical quality issues, allocating technology dollars and training employees to maintain quality for thousands of varieties of its Lipton tea, Cup-A-Soup, Wishbone salad dressing and Lawry's brands. "Several years ago we used to fill pages and pages of actual, physical notebooks with SPC calculations. We've since moved away from that and into automation," says Hare.

On-line SPC at Lipton's various U.S. sites typically includes at-line check weighing, where line workers put packages onto a balance, and automatic data collection mated to computer SPC applications takes over. Computerized SPC control charts are constantly updated with the new data. In some cases, the computer not only provides the charts and triggers alarms to signify set-point deviation, but displays process adjustment instructions. Process intervention at this point is manual because adjustments are relatively infrequent.

After the right level of automation is adapted to suit the right variables, thoughts turn to reaping the benefits. While everyone still seeks dollar payback figures, that's no longer where the ROI buck stops.

At Lipton, Hare says, "We're working [with SPC] to reduce variability, to build customer satisfaction and loyalty. You can't put a dollar figure on that, and yet look at how important and how valuable that is to an organization.

"It's important to realize that operators need an understanding of what is going on with control charts. It's dangerous to blindly tell operators to just follow these instructions.' So while the automation is an assist, it doesn't serve as a substitute for an understanding of what's behind a control chart and how it works. That's not to say that operators have to have a master's or Ph.D. in statistics to run control charts - far from it. But I think they need to have an understanding of the concepts of variability in the process."

How much data collection and SPC technology a process requires depends on the application. Some problems can be solved using common sense, while others require massive root cause analysis or number crunching. Sometimes, especially in large organizations where it's easy to replace logic with finger pointing, it's often best to let the numbers do the talking.

"You can't run SPC on every single variable," adds Hare. "Maybe a Du Pont has a need to run control charts on virtually everything, but in the food industry, we just don't have a need to apply that degree of engineering to all of our process parameters. Some variables are not really so much of a problem that they warrant the constant attention that a control chart would devote to the issue."

The level of technology

should suit the application

On-line sampling and manual data entry may be sufficient for a specific application, while dozens of PID loop controllers could be overkill both in terms of process control requirements and statistical signal processing. If common sense, ROI projections and budgetary considerations don't provide enough data to propose a quality program, you can always run a feasibility study using statistical methods.

"SPC makes you visualize what is the most economic point of control," says Hans Bajaria of Multiface quality consultants in Garden City, Mich. As SPC works its way through the organizational structure, it's the managers who must decide the appropriate level of data collection, including computer hardware and software purchases. And it's the plant floor personnel who are always: 1) the company's front-line quality defense and 2) the last to know what new so-called quality measure is about to be implemented.

"Now the operator has one more thing to do. If the [SPC] system is not convenient, if it is a burden, the operator will not use it. Or, he will sabotage the effort by faking the data," says Bajaria, who adds that this is a common practice driven by ignorance due to lack of training, not malice.

Explains Bajaria, "The most crucial point in SPC is that you make it very convenient. This is where the electronics and the software come in."

PC software explosion

While full-blown distributed control systems offer a real-time database, they typically off-load files to information systems. Similarly, programmable logic controllers offer highspeed processing of sensor data, but SPC information processing is handled by the supervisory computer.

While extensive historical files or real-time controller communications may prove a burden to garden variety desktop computers, the proliferation of low-priced personal computers and graphical environments like Microsoft's [Windows.sup.TM] have inspired a groundswell of SPC software offerings. More than 100 companies are marketing packages for the PC environment alone. (See accompanying story on page 26.)

"Several technological innovations are starting to cross. One is the communications technology between the PLC to the PC, which is widening from the traditional serial port to direct data highways. These new [communications] drivers allow a PC to grab data from the PLC backplane significantly faster than the old RS422 or 232 serial connections - it's like going from a country road to a freeway," says Tony Baer, a senior information market analyst with Advanced Manufacturing Research in Boston.

Have you bought a computer lately? Hard drives are getting smaller and store more data than ever before. Even with the currently inflated price of RAM chips, random access memory can be acquired for under 80 per megabyte. And with the coming of a new generation of microchips, near work station power will be available for little more than a few thousand dollars.

In addition to hardware considerations are "enablers like Windows," says Baer, "that allow you to do more things with your data, like interface with the spreadsheets, databases and report generation packages you already know and love.

"If you have a windows-based SQC or SPC application, you can feed that data into EXCEL [a software spreadsheet] and dynamically update your spreadsheets or graphics, or whatever windows-compliant package you have."

Of course, the whole Quality house can fall apart if, as Bajaria, Hare, Surak and others continuously remind us, operators aren't properly trained.

U.S. education 'deplorable'

"To emphasize technology as a solution is, in a way, to perpetuate the way American industry has been attacking problems in the past, and that is to throw technology at it," says Matthew Kline, whose Chicago-based firm is dedicated to SPC training. "I'm not saying technology's not important, but to emphasize technology and ignore the people is missing the boat.

"TQM and SPC are all about what people do, not about what technology does. Though there's some advantage to going to automated data collection, on the whole, when people are trained to do SPC on the floor, they do it manually. And that's what they should do, for at least six months, before automated systems are even introduced," says Kline, who feels that in cases where the data processing requirements are light, manual SPC can accomplish the same goals as automated systems, right where the problem is located - not across the plant floor in the process supervisor's office.

The major change in business from TQM and SPC, feels Kline, "is that people who have traditionally used their hands to do the job are all of a sudden utilizing their hearts and minds to solve problems."

But is the end result that management heaps more work on them, getting back the numerical equivalent of fudge? Questioned about the difficulty of asking unskilled or semi-skilled employees to participate in SPC efforts, Kline raised more than an eyebrow.

"The U.S. has a very poor public education system. Compared to Japan's, it's deplorable. The Japanese high school diploma is equivalent to a bachelor's degree in the U.S. And the German work force, because they have a great apprenticeship program, is also vastly superior to ours. What has to happen is that private industry has to compensate for what's lacking in public education. "When companies start investing in their people, that's where U.S. manufacturing can make real gains. The companies that put tools like SPC and TQM in their [people's] hands, and teach them statistical thinking and the need to reduce and eliminate variability ... these are the companies that are going to become world-class manufacturers."

What's in the SPC tool chest?

While you may not use all the available tools to build your house of quality, it's good to know you're free to choose from the seven basic tools of Statistical Process Control:

1) Flow charts are a starting point for simplifying and depicting all the steps of a process, value-added and otherwise.

2) Pareto diagrams measure problems against the frequency of occurrence and are helpful for setting problem-solving priorities.

3) Cause and effect diagrams, also called the fish bone or Ishikawa diagrams, help identify primary, secondary and tertiary causes of, for instance, process variability.

4) Histogram charts provide a frequency distribution picture of measurements over time and are often used with run charts or control charts to better understand trends over time.

5) Scatter diagrams depict the extent of variation between X and Y variables. They may indicate a correlation, but be careful not to draw erroneous cause-and-effect relationships.

6) Run charts demonstrate how a process performs over a specific time period, typically shorter than that of histograms. Individual values, even means and ranges, can be plotted over time.

7) Control charts are run charts with upper and lower control limits. These are ideal for tracing specific parameters once they are identified. There are several types of control charts to depict variation.

PC-based SPC software suppliers

Advanced Systems and

Design, Inc. Air Academy Press American Institute for Quality

& Reliability American Quality Systems, Inc. Applied Computer

Solutions, Inc. Applied Statistics, Inc. AT&T Manufacturing Systems Avair Software BBN Software Products BCR Computing Corp. Branson Untrosonics Corp. Brunswick Instrument Challenge Micro Systems Challenges International Cimnet Systems, Inc. Cimple Solutions Concert Corp. Concourse Corp. Philip Crosby Associates, Inc. The Crosby Company Cybermetrics Datamyte Division

Allen-Bradley Co. Datanet Technologies, Inc. Datonix Daytronic Corp. DES Systems Dimensional Data Systems Dynacomp, Inc. Federal Marketing

Services, Inc. Fowler Co., Inc. Gagetalker Corp. Gensym Corp. Geomet Systems, Inc. Ginesys Corp. H&H Servicco Corp. Hertzler Systems, Inc. Iconics, Inc. Infometrix, Inc. Intec Controls Corp. Integrated Technologies, Inc. Intellution, Inc. Interaction Research

Institute, Inc. Intercim Corp. International Quality Institute IQS, Inc. John A. Keane and Associates Labtronics, Inc. Lionheart Press, Inc. Major Micro Systems, Inc. Manugistics, Inc. Martin Software, D.A. MAS Software, Inc. Metristat Divisio, BSD, Inc. Metrscope Corp. Minitab, Inc. Navone Engineering NCSS Nimac Software Northwest Analytical, Inc. Nutek, Inc. Penton Software, Inc. Perry Johnson PHT Consulting Associates The Pister Group, Inc. PQ Systems, Inc. Process Control Software

Systems Process Design & Control Process Integrity Inc. Product Integrity Co. Productive Imaging Professional Applications

Development QA Training Tools Qualitron Systems, Inc. Quality Alert Institute Quality America, Inc. Quality Concepts, Ins. Quality Cornerstones Quality International, Ltd. Quality Measurement systems Quality Resources Quality Resources

International Rapid Data Research Information

Associates, Inc. SAS Institute, Inc. Scientific Computing

Associates Sector Systems Co., Inc. Slo-Town Systems Sof-Ware Tool Kit SPSS, Inc. The L.S. Starrett Co. Statimate Systems, Inc. Statistical Graphics Corp. Statscan, Inc. Statsoft, Inc. Stephen Computer Services,

Inc. Stochos, Inc. Strategy Plus, Inc. Syscon International, Inc. Syscon-Plantstar TA Engineering Co., Inc. Technical Center Courseware

Co., Inc. Teque, Inc. Teranet IA, Inc. Ultramax Corp. Verax Systems, Inc. Zontec, Inc.
COPYRIGHT 1993 Putman Media, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1993 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:statistical process control
Author:Sperber, Bob
Publication:Food Processing
Date:Oct 1, 1993
Previous Article:What's new in aseptics?
Next Article:Keep it clean! Plant layout and design are keys to sanitation maintenance.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |