On Friday, January 1, 2016 at 9:41:01 PM UTC-5, Tim R wrote:
I'm a mechanical engineer but I've been in management a long time, I don't get to do much technical stuff.

This week I got called to a gas leak. The emergency response and the mechanics had already been there and confirmed safe to reoccupy, but somebody wasn't so sure and called me.

These things are 80% one problem and 9% the other. The crew that showed up found and fixed the 80% problem, but it wasn't actually causing the leak.. The leak was somewhere else. Just because one thing is broken doesn't mean something else isn't broken too.



Once a problem is identified, attentional blindness sets in. Contrary evidence is not just ignored, it cannot be seen.

In this case the pilot light was out in the kitchen, and they assumed the gas smell came from it. They shut off the gas to the kitchen and authorized the building to be reoccupied.

They ignored the evidence the real leak was somewhere else - well, they were incapable of seeing that evidence because they already knew what was wrong.

When I got on scene it took little time to find where the smell was coming from. Partly that's from my bias to try to avoid assigning the cause until I've looked at all the evidence, but partly it's from knowing human nature and knowing if it were the usual problem they would have fixed it the first time, and it obviously wasn't.

I sent the mechanics back and told them where to look, and the second trip they fixed it correctly. Of course I had to keep quiet my part in it, there's always politics in these things.

So yes, fix the most likely fault first, but bear in mind the potential cost of it being something else.

I hear that. I remember being stuck on a problem so I asked a friend for a fresh look at it. He looks around then thought about it for a minute. He then looked at me and said " if you touch your head and it hurts and you touch you arm and it hurts and you touch your toe and it hurts maybe it is your finger that is the problem ". That was a strange way to put it but he was 100 percent right as I had my scope probe on 1 to 1 instead of 10 to 1 causing a loading problem. It is always a good idea to get a second opinion when you are spinning your wheels.

Another unrelated problem. You being an engineer will enjoy this one. The city has a water leak. They are not sure where the leak is. They can not dig up a whole city block to find it so they use a clever trick. They put two microphones about 50 feet apart where the leak seems to be. Then then record the random sound of water hissing out of the pipe. The two microphones are wired to left and right channel of the recorder. By adjusting the time delay of the right or left microphone they can find the sweet spot where the random noise from one microphone matches the random noise of the other microphone. Once the delay is know they know where the leak is within 1 or 2 feet. That is a clever trick.

There is a similar trick in cabling. You send a pulse down a cable that is about 80 percent c propagation speed. If you hear an echo you have a problem. By timing the echo at 80 percent c you know where the problem is.