Wednesday, March 8, 2017

A632.8.4.RB_BorutyAlan_Cynefin Framework Reflection

Cynefin Framework Reflection

The Cynefin model has 5 domains or contexts: simple, complicated, complex, chaotic, and disorder.  Disorder is where you are when you do not know what space you are in.  The normal response is to act according to your preferred method of action.  There is no decision model in this context, but the danger if you react to a situation based on your preferred method and it does not match the context of your action, you will most likely fail or even cause more harm (Snowden, 2010).

In the simple domain there is an obvious cause and effect that is understood by everyone and there is one right answer.  The model to facilitate decision-making is sense, categorize, respond.  Applying best practice is appropriate (Snowden, 2010).

Up until recently, water leaks at my sister’s house were a very common occurrence.  The piping that was originally installed overtime became brittle and weak.  The first leak flooded the place because the shut off at the house would not close.  That time I repaired five different leaks and replaced the shut off valve at the house.  I showed my sister and her daughter where the valve was and how to close it.  After the first repair I was under that house repairing leaks nearly once a week.  But, I never had to work under it in flood conditions again.  Unlike me, my sister’s hearing is excellent.  She would call and tell me she heard a leak and ask me if I would fix it.  She was right every time.

When I installed the shutoff valve and took the time to show them where the valve is, how to use it, and when to use it, I created a task for them that falls perfectly in the simple domain.  She hears water running (senses) she checks in the house to see if there is water running. If she does not find water running inside the house, she determines that it is a leak (categorize).  And immediately closes the shut off valve which stops the leak until I can get there to fix it (respond).  There is no ambiguity, cause equals effect, and repeating the pattern consistently produced the same result.  I replaced all of the lines a few weeks later and she has not had a leak, but if one does spring up, she will still know exactly what to do and it will produce the same results (Snowden, 2010). 

In the complicated domain, there is a relationship between cause and effect, but it is not easy for everyone to identify and it is possible that more than one right answer exists.  These issues require expert analysis.  It may be necessary to bring in experts.  The model to facilitate decision-making is sense, analyze, respond.  The use of good practice is appropriate (Snowden, 2010).

Not too long ago, I experienced an outage on my home phone.  I had seen telephone repairmen in the neighborhood all day, so I assumed they were working on the system and they would restore service by the end of the day.  When that did not happen, I checked with my neighbors.  They all had service and had not been without all day (sensed).  Then I assumed that I had somehow damaged the lines or an animal had chewed through the line.  I found that the line was buried over two feet deep and it was encased all the way to their side of the box.  I opened my side of the box, but did not see anything out of the ordinary.  At this point I had gathered up enough information to call the issue complicated.  I also knew that I still had the problem, but I had no idea what was causing it or how to fix it.  The extent of my analysis told me that I needed the experts to make any further progress (analyze).  Having exhausted my resources and capability, I called the phone company and put in a service request (respond) (Snowden, 2010).  They were at my house the next day.  I told them what I had done and found.  I also mentioned seeing one of their repair vehicles in the neighborhood the day I lost service.  They were certain that it was on my side and spent a few hours trying to find my problem.  Finally, they did what I could not do.  They opened their side of the box and found their line dead.  There is a panel box (not sure what it is called) just down the street.  On the day I lost service, a technician was removing service to a different house.  In doing so, he cut one of my lines.  He either cut the wrong line and forgot to reinstall it or broke it without realizing it when he was locating the one he needed to remove.  It does not matter, those things happen and I am certain it was not done purposely.  However as a technician, you always look at the last place people worked before the problem occurred.  It would have saved the phone company a half day of labor.  They found the problem within minutes and had it repaired in less than an hour. 

The cause and effect relationship became clear when they found the problem.  I feel that when I turned the problem over to the experts, my complicated problem should have become their simple problem.  They should have been able to use best practice from the simple domain.  That, of course, is just my opinion.

The complex domain is unpredictable and there is no apparent cause.  There are no apparent right answers.  Creative and innovative approaches are necessary; dissent and diversity are helpful (Snowden, 2010).  The decision model is probe, sense, respond (Snowden, 2010)..

I had just upgraded to the 5-skill level…and to be perfectly honest I had been signed off on a number of things I had not done yet.  Being a 5-level meant I could work on the system on my own.  I was a little nervous, but I had a safety net of qualified people I could question when necessary.  Two weeks later I was in a C-141 on my way to Saudi Arabia where I would be one of two Radar technicians.  There would be a 7-level on day shift and me on night shift.  My safety net was gone.  But I went in with my head up and a positive attitude.  Things went well for the first week.  Then an aircraft had to abort its mission because the main Radar computer called the RDC would simply turn off about every ten minutes or so, but it was very intermittent.  Without the RDC, nothing works.  The entire time it failed, it only generated a failure message one time.  The message indicated a power transient occurred.  So this issue started in the complicated domain because I had an idea what was causing the problem and I thought I knew how I would fix it.  On the aircraft I began taking power measurements.  The input to the cabinet goes through a filter on top of the cabinet.  I measured input and output of the filter and voltage and frequency were perfect and they were stable (probe).  The system failed but the readings did not fluctuate at all.  The same was true for every other power supply internal to the cabinet.  I continued for a few hours without finding a single problem from the front of the cabinet (probe).  It was time to access the backplane.  At that time, I had only been in the backplane once, in a class.  There are about 40,000 pins and 60 to 75,000 thirty gauge wires connecting all of the signals in the RDC to their location.  The way it is laid out, there is a real danger of shorting wires and/or breaking wires.  I really did not want to go in that night, but I had no choice.  When I opened the backplane, the failure went away, completely.  When I closed the door, it started failing again (probe).  In the cabinet there is a riser plate.  When I closed the door, the riser lifted the RDC chassis about a half an inch.  It dropped it by that much when I opened it.  It was consistent.  I had not repaired anything, in fact, I made it a little more confusing (sense).  But I could make the problem go away.  I spoke with the production superintendent and explained that if I got into the backplane the problem would probably get worse before it got better.  We had three aircraft and were responsible for airborne coverage 24 hours a day, 7 days a week.  Taking it down to two aircraft is not sustainable.  He understood and was good with the creative fix even if temporary.  The next morning we got approval to modify the cabinet (respond).  Removing the riser worked.  It did not fail again.

When the aircraft returned, the maintenance commander sat the aircraft down for maintenance.  I and another military Radar technician and two Westinghouse engineers were picked to work the problem.  We were to split up and work the aircraft around the clock until we fixed it.  That took three weeks.  After putting the riser back in the cabinet the failure returned.  We searched signal by signal for over two weeks and got nowhere.  We borrowed a piece of equipment from the mission computing shop.  It allowed us to probe 24 signals at the same time (probe).  Then we finally saw it.  The signal that protects the equipment for over current/voltage conditions changed state very quickly and very slightly, but it was enough to shut the RDC down.  Every input to that signal looked perfect except from the input filter.  It did not look bad it just was not as crisp as the others.  When we looked at it with a spectrum analyzer it was clear (sense).  The signal was noisy, but every few minutes the noise level spiked and the protective signal saw it as out of tolerance and shut the system down (sense).  Replacing the filter on top of the cabinet corrected the problem (respond).  I had never replaced one of those filters before and I never did again.

This context of this problem was complex we probed for weeks hoping to see anything emerge that we could follow.  It finally happened, but not until we employed the logic analyzer which allowed us to see relationships it was a first for radar and as a result we added it to our required equipment documents and added technical order procedures to use it.  Once we saw the protect signal change, we began to sense the problem was not a power problem, but noise on the riding on the power signal.  This was also the first time I had ever used a spectrum analyzer on power signals, but it worked and we responded by replacing the filter. 

The chaotic domain poses unique problems.  There is no clear cause and effect relationship.  There is no time to think and many decisions have to be made immediately.  The situation is highly stressful and turbulent.  It is pointless to look for the right answer instead you must act immediately to establish order.  The model to facilitate decision-making is act, sense, respond (Snowden, 2010).

This example affected me profoundly, but I was not there.  I will share the actions a friend of mine took to establish order in a chaotic situation.  My friends name is Craig, He was a Senior Master Sergeant at the time. 

On Elmendorf Air Force Base, Alaska at 0745 on September 22, 1995, Aircraft 77-0354 ingested several Canada geese into the number 1 and 2 engines on the left wing just after rotation.  The number 2 engine was completely destroyed and the number 1 engine was badly damaged and operating at 50 percent or less and was experiencing repeated compressor stalls.  The aircraft reached 250 feet, but it could not produce enough lift on the left side and was right wing high.  Approximately 2 minutes after take-off roll the aircraft crashed into a heavily wooded area nose first and right-wing high.  The aircraft came apart, exploded, and burned.  All twenty four crew members including two Canadian nationals perished.  All were brothers, three were good friends.  Maintenance personnel watched helplessly as the aircraft they just launched exploded into a fireball less than a mile away (FSF, 1996). 

My friend Craig was on duty, but he was not the production superintendent that morning.  He of course heard the radio traffic and the explosion.  He took immediate control.  He grabbed two Master Sergeants (act).  One he directed to collect all of the aircraft records and bring them to his office then have the planners lock down all maintenance software and automated systems (act).  The other Master Sergeant was to get the training records for every person and lock them in his office.  Then Craig went outside and directed the production superintendent to shut down all flight line maintenance and bring everyone inside the building (act).  Back inside Craig initiated a full recall of all personnel (act).  He also called the hospital and requested urinalysis support for the entire maintenance squadron (act).  He then called and requested Chaplain Support (act).  .  He found the Crew Chief for the aircraft and had a short, private conversation with him (act).  Then he talked to the entire group.  He explained all of his actions and why each one was necessary.  (sense).  He told everyone not to talk about the accident to anyone except Chaplains and spouses.  Until the investigation is complete everything is speculation.  He encouraged them to talk to each other and discouraged anyone from blaming themselves (respond).  The commander came in with a chaplain and Craig retired to his office where I assume he had to let things soak in (Brian, Eric, & Craig, 1995).(respond)

This topic is still difficult to discuss.  Twenty four Airmen died that day.  They were doing something they genuinely believed in.  Many lessons were learned that day, lessons that changed people forever.  I am one of those people.  The situation immediately after the crash was chaotic.  People were in shock, disbelief, and denial.  Some were angry. The Dedicated Crew Chief was sobbing.  Craig took action.  He said he did because he did not know what else to do.  He was upset that he did not know where to find the emergency action checklist.  I never saw their checklist, but I do not think he missed too much.  Craig’s actions from the beginning helped ensure the crash investigation team had current information when they needed it. He also snapped his Airmen out of it so to speak.  He had established order by the time the commander came in.  I believe he got them into the complex domain by the time the commander arrived.  

Brian, Eric, and Craig were asked and agreed to go out to the crash site to search for aircraft pieces and debris…none of them will discuss that experience.

References:

Snowden, D. J., (2010). The Cynefin Framework. YouTube. Accessed 2017 February 28. Retrieved from:https://www.youtube.com/watch?v=N7oz366X0-8 (Links to an external site.)

Flight Safety Foundation (FSF) Accident Prevention, Volume 53, Number11, November 1996. Retrieved from: https://flightsafety.org/ap/ap_nov96.pdf


Conversations with Brian, Eric, & Craig, (1995).  True to the best of my memory.

No comments:

Post a Comment