Strategies for Industrial Maintenance, Operations, and Engineering

This post is the third and last entry in a short series on interviewing candidates for knowledge and skills in reliability. See also Part 1: Root Cause Analysis and Part 2: Bearings. With these three tools, a fairly comprehensive picture of a candidate’s knowledge and skills can be gleaned in a short time.

Question:

Tell me about a time that you analyzed a system or piece of equipment for failures and then designed a reliability strategy.

Follow-up Questions:

• What process did you use?
• How did you get the recommendations implemented?

Comments:

• RCM is the flip side to RCA: it is a kind of “RCA in advance.” It is one of the most challenging task that a reliability professional will undertake. Not only do team RCM efforts tend by their very nature to constantly go off track, creating challenges for the facilitator, but the facilitator must direct the group to the correct level of detail: deep enough, but not too deep. Often, the facilitator has to overcome the perception that it is a waste of time. Finally, the recommendations generally span multiple departments and RCM facilitators have to “influence without authority.”
• The RCM process is surprisingly controversial in the reliability world. It has a reputation for “getting in the weeds” and not creating enough value to justify the expenditure of manpower, but that can be the result of inadequate facilitation, an improper team ordering, or inadequate preparation. My personal feeling is that RCM is a tool that is best applied once a general reliability program including lubrication, inspection, and condition monitoring has been established. Otherwise, for each piece of equipment there will be dozens of recommendations that could be addressed by general efforts at establishing the “basics” at a much lower cost in manpower.

Qualities of an Ideal Answer

• Before jumping straight to recommendations, the process should spend at least some time on system and equipment function, functional failure, and FMEA, addressing failure effects and consequences separately before deciding on tasks. Ideally the cost of periodic tasks, whether on the run or during shutdowns, should be weighed against operational consequences unless there is a safety or compliance aspect.
• It would be best to have an experienced and credible reliability engineer sit through this question. However, many of the comments regarding root cause analysis, above, also apply here, particularly those on driving organizational change.

As promised, here is the first in a short series of posts on interviewing people for reliability positions. These questions will fit well in a behavior-style interview for evaluating knowledge and skills.

Question:

Tell me about a time when you had to determine the causes of a recurring problem.

Follow-up Questions:

• How did you determine the causes?
• What were the solutions proposed?
• What were the results?

Comments:

• Since this is not an initiative question, the candidate does not need to show that they proactively identified a problem. Either they could have found the problem themselves or it could have been given to them. It would actually be best if the problem was given to them since it is less likely to be a problem which they have chosen as particularly interesting or which they have special knowledge about. This helps isolate K&S factors from initiative.
• The plural form of “cause” is used on purpose. It is very unlikely that any problem tracked down to latent roots has just one cause. Therefore, any response that settles on just one cause is unlikely to be as thorough as RCA actually needs to be. An alternative approach would be to ask about a singular cause, in order to try to misdirect them to think it is okay to identify just one single cause, but it might then get into the realm of “trick questions.” The phrasing clearly indicates the expectation that more than one cause is identified.
• Portfolios are under-used in the technical professions, but the ideal candidate will bring written examples of their RCAs in preparation for questions like this. If the printed example follows a logical structured process, they are probably in the top 1%. However, probe to ensure that the example has some substance behind it, and question anything that looks like an assumption.

Qualities of an Ideal Answer

• The failure mode is well defined. Sometimes, the failure definition will change midstream as new information becomes available and the problem is not what people first thought. That’s good. Sporadic problems are generally just what they appear to be: “The shaft broke and the roll fell.” With chronic problems the definition of the failure mode often changes as new information is gathered. The answer should sound almost like a story.
• Drives to multiple latent causes. Stopping at physical cause not acceptable as a failure analysis. Examples of physical causes are: lubrication breakdown, shock load, fatigue, contamination, and abrasion. Stopping at human cause (decision errors) is also not acceptable unless it was proven to be an intended violation or act of sabotage (rare). Examples of human causes are: decision not to follow procedure, decision to use different lubricant than specified, decision to fill tank all the way instead of ¾ full, decision to use screwdriver as pry bar, and decision to use hammer instead of induction heater to install bearing. Latent causes come from asking “why” about human causes, and often deal with perverse incentives. Examples of latent causes are: buyer’s performance measured by cost savings, mechanic received frequent past recognition for getting the job done quickly, and correct tools/parts/supplies not available or out of stock.
• Verification used at all steps. No assumptions are made at any point in the analysis based on past history. Human and latent causes are verified through interviews and feedback from people involved.  Strong interview skills are a must for RCA work. Good interviewers have to “hear” what other’s don’t say (whether due to embarrassment or malicious intent) as much as what they do say.
• The solutions proposed tie directly back to cause. Solutions must not be detached from cause. Proposals can address any of the causes, not just “root” causes, but a direct link must always exist. The department that owns the cause (Engineering, Operations, Maintenance, or Procurement) must own the solution. For example, an equipment installation problem must not be addressed with periodic maintenance inspections.
• Demonstrates economic thinking in the level of analysis chosen. Example: metallurgical analysis is avoided if visual inspection can do the job. Metallurgical analysis would generally be chosen to answer a specific question whose answer can’t be obtained from more economic options.
• Demonstrates organizational savvy in driving change. Corrective actions should be reviewed and aligned upon by department managers before assigning and trying to implement on the floor. Convincing a department manager that a problem is coming from their department and winning their support in correcting it takes a level of tact and skill that will only be found in an exceptional candidate.