From golden batches to golden results
By Kevin Trantham
As lean manufacturing concepts coupled with six sigma initiatives have extended their influence into chemical batch operations, we often see a variety of attempts to continuously improve through the application of ideal or “golden batches.” The concept itself is simple enough and typically follows a “define, measure, analyze, improve, and control” process. However, implementing such a solution does not require six sigma training.
For example, consider the single attribute of cycle times. The ideal batch is first defined with optimal cycle time durations being specified for each phase. Second, performance is measured on each batch relative to the golden batch standard. This is followed by some type of data analysis, which can range from random operator notes to fully automated data dumps of process conditions to integrated overall equipment effectiveness calculations. From there, the process can get fuzzy, as does the effectiveness of the improvement and control.
Good data is the foundation for success
To reduce this fuzziness and improve success rates, we begin by looking at key characteristics in the “define” and “measure” steps required to produce good data. Good data alone will not produce continuous improvement; the proper framework that establishes ownership, operational rigor, and accountability must also be in place to maintain control and drive continuous improvement.
To define good data, first consider the following common data problems:
- The data collected is not granular enough.
- The data collected is not consistent.
- Collection is on a “voluntary” basis, as the actual data and the data reported by cause differ greatly.
The collection of data is overly intrusive to production. Capturing data should require fewer than five seconds of operator time and, if possible, be automatic.
When establishing how data is measured and collected for analysis, the implemented solution must remove these limitations. This can be achieved in a typical golden batch application by first developing a single screen that alerts the operator to deviations from the golden batch standard. When a warning requires root-cause identification, the operator is then presented with a top-tier pull-down menu with categories including planned maintenance, planned wait time, raw material, and equipment malfunction. Selecting one of no more than 12 categories from a pull-down menu, the operator is then presented with a subcategory list of root causes (again, no more than 12) to choose among. So with two clicks from a standard root-cause tree, the operator can capture predefined root causes in fewer than five seconds and use the data for consistent, quantitative analysis.
Proper framework, combined with good data, drives success
The most common challenge in producing golden results is a framework that establishes ownership, operational rigor, and accountability:
Once we have created our root-cause tree identifying the possible detractors from optimal production, each root cause is then assigned to an individual root-cause owner (RCO). The RCO is empowered to cost-effectively improve performance for his or her root cause. To assess the cost-effectiveness of the solution, RCOs must be made aware of the cost (in dollars) of the detraction from optimal performance. In addition to having real-time dashboard awareness of this cost, RCOs should automatically receive a report of the cumulative cost of their detractor items at the end of each shift.
RCOs should be required to assess the daily costs associated with their areas of responsibility and identify improvements to eliminate deficient performance. Organizations can also establish significant exception events requiring immediate notification. Weekly, the cumulative contributors should be summarized and a report automatically generated to the RCO managers. RCOs of the top contributors to lost performance should meet with management to drive improvement.
When ownership with empowerment has been assigned to identify and implement solutions when they are economically feasible, the weekly management meeting improves dramatically—from reading a report of detractors to discussing what behaviors have produced better results. By establishing accountability, a management team creates a solution-driven environment.
With increased awareness from solid reporting of meaningful data, we often see peer-to-peer accountability drive performance as much as top-down accountability. As a result, golden batches properly implemented can produce golden results when the best know-how in process control and optimization is applied to support a framework for continuous improvement. Typically, there is significant low-hanging fruit where low-cost, high-return changes can produce the best outcome. Attacking operator wait time, optimizing heating/cooling, or implementing appropriate advanced regulatory controls can efficiently drive performance.