To deal with this, BASF has developed a process called “learning from others” that aims to ensure knowledge is systematically transferred among operational staff.
The process begins with a demographic analysis of operating staff and line managers identifying crucial knowledge carriers. A kickoff meeting then brings a person with expertise together with other relevant colleagues and the assigned successor.
“Key to this initial meeting is for the skilled veteran to provide a structured description of the knowledge relevant to the operational tasks involved. An interim report that captures the details then is presented to the plant management team,” notes Stautmeister.
Based on the report, the management team decides upon measures necessary to implement to achieve efficient knowledge transfer. To do this, the team considers questions such as: is the knowledge of the employee sufficiently documented; how can the documentation be developed to ensure that it’s complete, understandable and accessible and — if not — what recommendations to give on how to improve; which business-relevant practical knowledge of the employee should be spread around more employees; and are there means to improve processes and procedures?
Meanwhile, the Center For Operator Performance (COP), Dayton, Ohio, an industry/university alliance, continues to pursue research into many aspects of operator training. One key area is the use of its ShadowBox technique — this enables trainees when they are presented with decision points to benefit from the knowledge of subject matter experts (Figure 1).
Last summer, for example, a pilot was conducted on a fluid catalytic cracker (FCC) in the U.S. Both experienced operators and novices were run through different scenarios.
“We got what we wanted: we identified the difference in the level of knowledge between the two groups. Interestingly, none of the senior operators got the scenarios 100% correct, so even they were missing things. So it became an opportunity for them to learn, too,” says Dave Strobhar, a founder of the COP and chief human factors engineer with Beville Engineering, also in Dayton.
For COP members as a whole, two questions emerged from the study. First was whether or not a firm that has high-fidelity simulation available would be able to use it with ShadowBox to gain more value.
“So we put in decision points in the simulators to find out. Those who had experienced the ShadowBox approach performed 28% better in terms of diagnosing the causes of faults than those that didn’t. That was a very surprising figure for us,” adds Strobhar. The particular company involved is so impressed that it’s currently incorporating ShadowBox into its existing simulator training program, while also creating a users’ guide to help other companies develop and administer their own training scenarios.
The second question to emerge from the FCC pilot was about whether a scenario must be specific to a particular unit or can be more generalized — i.e., will each different FCC unit, for example, need its own ShadowBox or will scenarios from one FCC work on every other one?
“To test this, we are working with a company that has three polyethylene units which are similar but not identical. So we are looking into this and also into how the operators of one unit tackle decision-making on the other two units. The goal is to determine the value of ShadowBox training if only carried out on a similar (not identical) unit,” notes Strobhar. The results from this study should be released shortly.
Another issue the COP is looking into is low-fidelity simulation. “Many of our members prefer high-fidelity because it does mimic the real unit, but COP research has already found that low-fidelity is far superior in some situations. This is especially true at the beginning of the training process when high-fidelity simulation simply produces too much information. So another question we are considering is, ‘How much fidelity do you really need?’”
Although this question has yet to be tackled directly, at the end of 2016 the COP began a project to examine if low-fidelity simulation plus ShadowBox can benefit both. These results are due out in summer 2017.
Another concept that has captured the COP’s interest is that of mental models and how to improve those of operators.