Palmer Foundry Applies Quality Analytics to its Business

Reading time: min

After using a historian, statistical process control and enterprise manufacturing intelligence to improve product quality—and seeing results—the company is now turning the software toward internal operations.

Palmer Foundry, a maker of aluminum castings for semiconductor, automotive, energy, food processing, aerospace and other industries, has seen its fair share of winter weather at its plant in Western Massachusetts. But during a wicked bad storm a few years ago—when a power outage lasted days—Palmer Foundry had to pay careful attention to how its equipment was running on the backup generator. Specifically, the equipment that keeps the metal melted.

“If it’s not liquid, you get a two thousand pound aluminum lollipop,” said Bob Logan, president of Palmer Foundry. “And the crucible it is in becomes scrap. You can see $40,000 evaporate before your eyes in a power outage.”

So, they started up the emergency generator, thinking all systems were go. But they still ended up with big metal lollipops, with no explanation of what went wrong. To get to the bottom of this expensive mishap, they used an industrial process historian to identify where the process went out of control. By seeing where things were trending out, they were able to identify a 14-second power interruption that shut down the gas-fired furnace. And, they identified that the problem stemmed from an orifice on the fuel supply to the generator, which was undersized.

The Industrial Process Historian Database from Wonderware provided immediate insight to help the team diagnose the issue. “It’s like having an instant replay to look at the tape,” said Logan, an engineer, who, after graduating from Yale, played in the NHL for the Buffalo Sabres and Los Angeles Kings before acquiring Palmer Foundry with his brothers almost 20 years ago.

While Wonderware Historian keeps track of the time-series equipment data, another system in use at the foundry, called Quality Analyst from Northwest Analytics, is responsible for tracking and analyzing the transactional information of the raw material, process and finished goods

Palmer Foundry first found the Quality Analyst statistical process control (SPC) software several years ago when the company was urged by a semiconductor customer to become ISO 9000 compliant for quality assurance. As reported by Automation World in 2014, the Quality Analyst software was used to provide data charting and analysis of the casting process, which must include specific percentages of constituent elements, such as silicon, titanium, copper and magnesium, to meet quality standards. The company also invested in Northwest Analytics’ Focus Enterprise Manufacturing Intelligence (EMI) software, which provided another layer of data detail.

At that point, the goal was to show customers that Palmer Foundry understood the key process parameters required to ensure good castings. The software pulls from different data sources, such as raw materials and lab results, and alerts managers when something is trending out of control. “It is like a heart monitor keeping a pulse on the process,” said Jim Lagrant, vice president of engineering for Palmer Foundry. But it quickly proved even more valuable by identifying external issues, as well. For example, on one occasion last year, the casting chemistry was out of spec. The Palmer Foundry team was quickly able to determine the problem was from a raw material supplier. What normally would have taken a few weeks to figure out, took one day.

The data charts give the company the ability to have a completely objective conversation with customers. What once was finger-pointing is now proof points. “Once you show the control chart that tracks over time, you are having a different conversation,” Lagrant said.

Taking Care of Business

Palmer Foundry has had so much success with the SPC and EMI software that it is now applying it to its business, from training to safety to on-time delivery of product. By looking at business relationships—vs. processes—the company can see how it’s performing overall and where it’s going.

“Business is a process,” Logan said. “When you can measure something, you can manage it. If you don’t have a baseline measurement, how do you know if you’ve made improvements or not?”

Palmer Foundry’s new business performance system will be the focus of a presentation by Lagrant at The Automation Conference taking place May 22-23 in Chicago. Lagrant has been able to take an operation that is not highly automated and position it for a digital future by connecting production operations and the enterprise business functions.

Using first pass quality KPIs on molding lines, the plant and the entire company as a whole, Lagrant has visibility into operations to improve internal processes so that no aspect of quality is ignored.

For example, KPI charts provided to plant supervisors and upper management show monthly and year-on-year casting conformity from one of the company’s green sand molding lines. The charts update whenever new data is entered. In September of last year, there is an outlier, which was due to a new operator who was then re-assigned to another process, after which the quality metric rebounded in October.

“Although there was a trend down in the last two months of 2017, the overall average coming off of that line was 98.6%, which is over our target of 97.5%,” Lagrant said. “Having the top-level view of the data prevented us from overreacting and assigning our limited resources to solve what is in essence a small issue.”

Moving forward, Lagrant would like to give the analytics tools to operators in the plant so that they can see the performance of their work centers, or how the chemistry elements are reacting. “We would give them the tools to make good decisions…and boost results right away,” Lagrant said.

The only barrier to SPC and EMI on the plant floor is operator training, so that everyone would have a clear understanding of what a chart looks like when something is trending out of control. It is just an educational effort not operator resistance, because the plant floor team already understands the significance of the software. “By showing control charts to operators in our melt department, we were able to show how their actions were affecting the chemistry,” Lagrant said. “We were able to drive more stability into the process because the operator changed his actions.”

Alternatively, empowering operators could be as simple as creating alarms, so that they know there is a change in the process and some action needs to be taken.

Ultimately, applying the software to the business is about making informed decisions. “We have to be a better supplier to our customers,” Logan said, and that starts in-house. “You’re only as good as your last shift.”