Failure for an Engineers

Lately, I was thinking about "Failure".

Many scholars in the past have written many quotes on Failure.

" Failure is key to success "

is the most famous one, I guess. Some say

"Do not be embarrassed by your failures, learn from them and start again."

etc etc... It's one of the most subjective topics. Failure for a kid maybe not be able to buy a toy, for high school guy maybe scoring fewer marks in a subject, being a UG student it may be failing to get a placement or backlog and the list goes on.

But Being an Engineer or a doctor a small fault may be fatal for someone. I don't know about doctors but being an engineer Its very crucial to keep all the possible failures in mind while we design anything and try to prevent them. I remember once our college teacher said if you are safe after an accident, before thanking God thank the engineer who had designed the vehicle to keep you safe for such an incident. One of my favourite quotes on failure is

I have not failed. I've just found 10,000 ways that won't work.

From my small experience working on Formula student EV, We do Failure Mode and Effect Analysis aka FMEA Report on our Powertrain design. Let's discuss the different components of the FMEA report and how one should approach it. (It's applicable for any system but here I will discuss based on FS EV.)

1. Component/Item
2. Function
3. Failure Mode
4. Failure Cause
5. Failure Effect ( Local & Global )
6. Severity (Rating & Reasoning)
7. Occurrence (Rating & Reasoning)
8. Failure Detection (Rating & Reasoning)
9. Risk Rating
10. Failure Handling ( Vehicle and Team )

I'll discuss all of the above points briefly. So that you know how to approach the FMEA.

1. Component/Item - It is the most obvious and self-explanatory point. It's the particular component on which you are going to do the analysis. It is a small part of a subsystem.
2. Function - It's also pretty much self-explanatory. Here we have to mention the function of the component on interest.
3. Failure mode - It specifies the means of failure, what do we observe as a signal that the failure of the particular component has occurred. There can be multiple modes for a component to fail. Different components may have different or similar modes depending on their specifications.
4. Failure Cause - As we know a particular Mode of failure for a component, then it's time to think for the probable causes. When we analyse it we have to consider all forms of causes mainly Mechanical and Circuital. There are generally multiple causes for a single mode. We should try to consider all the possible ways the component can fail.
5. Failure Effect ( Local & Global ) - It is also self-explanatory. here we mention the effects of the failure. Local effects mean the impact on the region around the component and its effect on the particular subsystem to which the component belongs. Whereas the Global effects mean the impact on the whole system( EV in our case ). While writing the effects make sure it is in line with the modes, cause( mechanical causes have a different effect from circuital ones) and component function.
6. Severity (Rating & Reasoning) - Now if the failure occurs then what will it damage( just component or driver/operator ), how much, and are there any safety measures designed to reduce the severity. These are mentioned in the Severity. We use a rating system on a scale of 1 to 10. The higher possibility of damage to the driver/operator, the higher is the rating. Subsequently, we have to give the reasoning for our rated value. It should be in line with the effects.
7. Occurrence (Rating & Reasoning) - Here we have to mention the probability of the component failing. How often the component can fail? what do we do to reduce the failure possibility? Such questions need to be answered when it comes to Occurrence, or simply how we have designed to prevent the failure to occur in the first place. We use a similar scale as mentioned in severity for occurrence rating. The higher the possibility of the component failing, the higher will be the rating. Subsequently, we have to give the reasoning for our rated value. It should be in line with the failure mode and cause( a mech cause may be more likely to happen than the circuital one ).
8. Failure Detection (Rating & Reasoning)- There are various detection systems in place to avoid the small failures turning into disasters. Here we have to mention how likely the failure can be detected. Easier it is to detect the failure lower will be its rating on a scale of 1 to 10. Of course, you have to mention the failure detection rating reasoning. It should be again in line with mode and cause.
9. Risk Rating - the total risk is calculated using a simple formula i.e. risk rating = severity rating * occurrence rating * detection rating. The higher the rating, the higher is the risk for the particular component to fail.
10. Failure Handling ( Vehicle and Team ) - And finally when the failure has occurred how are we dealing with it? how are we( or the vehicle itself ) curing it? if there is an inbuilt system or mechanism installed in the vehicle to react to the failure, immediately after it occurs, it comes under the vehicle handling. After a failure, the team has to operate( wearing proper PPE kits, it needs to be mentioned as well ) on the vehicle in order to repair or replace the component or work on the cause to avoid it. These come under the team handling part.

It's not always possible to avoid a failure by 100%, there will always be some chance of failure to occur. The important thing is to study the failure properly in order to design a safe system. With more test cases we can add new failure modes, to make our system even safer. As Engineers, we have to think and prepare for the worst.