Inside View

Pennsville: Continuous Improvement (CI) Program

Fishbone diagram used to brainstorm causes of a problem

Recorded delays in production of this process since 2014

Dryer cycle time chart showing impact of unload timing change

Top row from left: Mike Widmer, Sam Midili, Craig Gunnett, John Keenan, Don Twigg. Bottom row from left: Chris Hood, Stephanie DiDonna-Krough, Barb Vergara

In February of 2017, I was chosen to participate in Siegfried’s Green Belt Program where I would learn new skill sets for analyzing and solving problems.

To earn my green belt, I was assigned to lead the project of ensuring that the Pennsville site would manufacture and release 35MT of our largest exclusive product. We have made this product for several years, but not consistently, and never at a sustained high rate. Historically, we had proven we could run at an average efficiency rate of 85% for a few months at a time, so the goal was to eliminate the 15% of inefficiency and prolong that rate for the entire 8 months.

We held weekly meetings to discuss potential improvements, give updates on active tasks and to reaffirm each week that this was the #1 priority for the site. We used like Short Interval Control, Dashboards and Process Mapping to name a few. We did a lot of data mining by using graphs and spreadsheets to see trends, historical values, look for wastes and to find areas for improvement.

We used the DMAIC model (Define, Measure, Analyze, Improve, Control) to organize our projects.

Phase D – “Define the project”
Producing and invoicing 35MT of this product in 2017 was already our known goal prior to starting the project as it came out of the company-wide goals for the year.

Phase M – “Measure potential drivers”
In Phase M, you define the influence factors to determine what areas of concentration are needed to see the changes you want.

A fishbone diagram is used to find causes for a problem. It’s a way to brainstorm and arrange results, then you can prioritize the problems you want to resolve. In ours (as shown below), we knew the problem was “sustaining cycle time”.

Once the causes were identified, we chose the ones we expected to have the biggest influence on cycle time and reducing delays and made them our main focus. We identified also other areas that weren’t as obvious to the group in the early stages such as the frequency of certain samples failing or the staffing in QC.

Phase A – “Analyze potential drivers”
Phase A of the DMAIC process is where you confirm the improvement potential for the changes you are making. This is typically done by measuring results and plotting them in a graph or chart.

In a graph, we plotted our actual dry train cycle times in real-time and we could see how the improvements made were getting us back on track if there was a slip.

Phase I – “Improve and implement”
During Phase I we were able to implement multiple smaller improvements that added up to a great benefit on the process. Some of these changes were: piggybacking batches slightly differently, putting a 3rd acetone trailer into service, unloading the dryer before cooling to 25°C, ordering backup parts for equipment prone to failure, prioritizing labor throughout operations to keep this product running continuously and returning QC staffing to 24x7 coverage.

Especially we introduced a clear instruction how to unload the dryer. The practice of waiting to cool to 25°C was found to be an example of waste. This would save 1–2 hours / batch since we wouldn’t have to wait as long for the dryer to cool.

Phase C – “Control the improvements”
In the Control phase we lock in any changes made to ensure they are always done in the future. Think of the continuous improvement process as rolling a boulder up a mountain. Each time a change or improvement is made, you stick a wedge under the boulder to make sure it doesn’t roll back down. Those wedges are the controls. A lot of the controls we put into place are captured in updates to the batch log records. Setting up SAP item numbers for maintenance parts and establishing a min/max level in inventory is another control we agreed to which will prevent us from stocking out of parts that are needed to minimize downtime.

Mission Accomplished
By the time we wrapped up the project, we had eclipsed our goal of 35MT. By year’s end we had produced and sold over 46MT of commercial product. We averaged 51.7 hours / batch which was better than our historical average of 55.2hours / batch. And we were able to realize additional gains to the process by a combination of higher yields, faster internal and external QC testing and reduced downtime compared to previous campaigns.

One of the biggest drivers of our focus was the use and distribution of a dashboard, that summarized our key KPIs for the project, which was reported up to leadership and posted in the wet trains and dryer building.

But the main reason we were able to achieve and exceed our goal was due to the people involved in seeing these efforts through. My project team consisted of John Keenan, Sam Midili, Craig Gunnett, Janet Nixon, Stephanie DiDonna-Krough, Barb Vergara, Mike Wid- mer and our sponsor Chris Hood. It took the full cooperation and dedication of Sales, Purchasing, Warehouse, Planning, Process Engineering, Wet Trains Operators, Dryer Operators, Shift Supervisors, Building Owners, Management, Maintenance, QC and QA. It proved just how effective we can be as a team.

I learned many new tools through the Green Belt program and had excellent support from Marcel Fischer on anything related to the project. To manage a project like this it requires a substantial amount of time and resources, so without the dedication from yourself, your fellow teammates and support from management, it would be nearly impossible. The Green Belt program may have given us the tools, but they only work if you have the right people using them.