Software provides better access to SPC data

The worldwide leader in premium pet foods uses statistical process control and continuous process improvement at every stage of its production operation.

When it comes to feeding cats and dogs, the Iams company is serious about quality.

The worldwide leader in premium pet foods uses statistical process control (SPC) and continuous process improvement at every stage of its production operation.

"Our quality system is developed around the whole production cycle", says Dave Bechtel, Iams director of worldwide quality.

"It entails every phase of the manufacturing process: from development to procurement to manufacturing to shipping and distribution to customer follow-up".

The Iams processing plant in Aurora, Nebraska, is ground zero for the company's quality and process improvement efforts.

New SPC programs were first tested and implemented here, then rolled out to other plants in the enterprise.

A typical Iams facility, the Aurora plant produces 200,000 tons of premium dry cat and dog food each year.

It employs 150 people in raw material receiving, milling, processing, packaging, warehousing and shipping.

The production process begins with testing ingredients to make sure they conform to specifications.

Then the ingredients are placed into bins before being put into mixers.

From the mixer they are placed in an extruder and then cut into the sizes and shapes demanded for various products.

After cutting, the pieces go into a dryer, where the moisture is reduced to specified levels.

External fat and ingredients are then sprayed on and the pieces are cooled and packaged.

At each step, Iams uses SPC to determine whether the process is capable of meeting specifications (in statistical terms, this is expressed as Cpk).

Once assured that it can, operators gather routine day-to-day data for trend analysis.

From there, SPC efforts focus on establishing tighter and tighter controls to decrease variability.

"We know we're making progress when line operators are no longer running to specifications, but to targets that were established by our on-site manufacturing process control teams", says Clint Paisley, quality assurance manager at Iams Aurora.

A good example of how Iams uses SPC, is to more closely target the accuracy of package fill weights.

As all packagers know, underfilling is not an option, both because of customer satisfaction and strict regulatory requirements.

Overfilling is one way to make sure requirements are met, but is bad for the bottom line because it adds cost.

The goal is to get as close as possible to the target fill weight without going under.

During each shift, an Iams operator measures fill weights and plots the data.

At shift's end the data is entered into a statistical software package, NWA Quality Analyst.

The software analyses the data and when a new shift starts, gives the incoming operator a statistical profile for the preceding shift: average package weight, the number of measurements below the lower control limit, and the number that exceed the upper control limit.

"We provide operators with the information and, if they see that a particular product is running outside of limits, the process needs to be evaluated", says Paisley.

"We look to run on both sides of the target with minimal variation and not under declared value weight".

Years ago, before the statistical process control programme was implemented, operators didn't have the data to evaluate capability from one filler to another.

"Without data, we couldn't optimise each line for fill weights", says Bechtel.

"Now we've got it defined and have narrowed the targets substantially".

Bechtel says that NWA Quality Analyst enables them to take data and - from a capability standpoint - determine what their systems are able to deliver.

"For a long time we used a standard fill weight target", he says.

"Now, with NWA allowing us to talk specifically in terms of data instead of in generalities, we've been able to optimise fill weights, generating savings while staying fully compliant.

And, as we get more involved, we're tapping into NWA's more technical capabilities such as t-tests and regression analysis.

There's room to grow as our quality program develops".

Iams chose NWA Quality Analyst for SPC data analysis because it is exceptionally easy to use.

"I'd been using Excel for SPC, but found it didn't provide the desired statistics.

With the amount of data generated, the files became much too large to handle", says Paisley.

"I wanted software that was easy enough that the operators would be willing to use it for decision-making.

NWA has all the SPC tools I want and it has exactly the easy-to-use interface I need".

Using the software, Iams operators can see how the process is performing by using I/R charts and histograms.

The software makes it obvious when rule violations (instances in which something in the process is outside normal parameters) happen.

When rule violations occur, operators and supervisors record the causes and corrective actions in log books.

Paisley uses the information to track changes to control limits and machine maintenance.

"With NWA, people no longer have to depend on gut-feelings or guesses about how the process is performing", says Paisley.

"They have real data upon which to make analyses and then decisions.

This leads to high consistency".

Before Iams started using NWA software, most of the quality data were kept in Excel spreadsheets.

"We had pages of data that was not being manipulated in a way that was helpful", says Bechtel.

"We were able to get averages, but weren't able to talk much about variability and standard deviation.

We didn't have a glimpse of the whole situation.

Now we're applying analysis to many product characteristics, using NWA as an assessment tool".

Paisley says the biggest value of NWA is its ease of use.

"Other software programs just spit out stats that intimidate people", he says.

"We don't need complex statistics.

We need basic tools such as how to collect data, how to look at standard deviations and simple work instructions on situations that occur.

Today, I have operators who can readily pull up histograms and use them to make the process better.

NWA Quality Analyst is perfect for that".

NWA Quality Analyst is supplied and supported in the Uk and Ireland by Adept Scientific.

Back to top