1. So basically my goal should be to take into account all significant contributor into my uncertainty budget (and then add them to find Expanded Uncertainty). If I want to be contain 68% of the data I take standard deviation. If I want to contain 95% of data. I multiply standard deviation by 2 and so on.
Yes. 95% seems to be the most common expression. However, it is up to you, and what level of confidence you wish to have.
2. I could only think of three variations that contribute significantly to my verall uncertainty budget. They are Equipment Error (Repeatabilty), Operator (Reproducibility), and Calibration. We developed Fish Bone Diagram to find the causes that would lead us to inaccurate measurements and all the other causes (30 or so) seems to contribute not as much as these three. Do we have to include them into Uncertainty budget?
Good question. I would say yes, but there are no values assigned. You would document that you considered a particular variable, but given XXX reason or whatever, it was considered insignificant. That way, if someone ever examined your budget, all the considerations have been accounted for. Too, having that listed may help contribute to a better budget. Someone may look over it and have some better information to contribute. All the information there together helps that process.
3. If we do include into the uncertainty budget. What are the different ways to calculate the uncertaity due to those specific causes that are hard to define and apply any statistical method to.
Of the 3 you have listed, they should be pretty easily estimated. However, you have asked the million dollar question.

One way that will work is to determine the variable and estimated based on the parameters. So say you are talking about vibration. You have measured the vibration, so you contact the mfg. and ask about how much vibration can effect the reading. They may can give you a value within a range or something that it would contribute.
The best way? Control or eliminate those variables where they are not of concern.

Like temperature and humidity (for some equipment). Find out from the mfg. what range the T/RH can be within without affecting the equipment's performance, and control it where it is not a factor.
As you are building your budget, put everything in the same designation. So have everything expressed in microns, diopters, degrees C, etc. You don't want to mix things in your budget; so convert them as best you can.
*******
There are many here who are way better versed in uncertainty budget than I. What I have found though, is this is a much better way to approach measurement systems. The budget gives you a start, and you begin with your "best assessment" of the measurement system you have. And... it will change. The values will change, and possibly other considerations might need to be considered. Also, you can lower your uncertainty (if needed; desired) by gaining better control over some of the contributors in the budget.
