For (3) Vince Adama book is available at Amazon. "Building Better Products with Finite Element Analysis". That is a good starter. Also Roark and Young has a chapter on failure modes that should be required reading.
For (2) Glad you asked that question as it gets past the Red is Bad, Blue is Good mentality. Always remember, FEA gives insight, not answers.
For (1) I can't really tell you how it is going to fail based on the information you provided. But I can ask some more questions for clarification.
Let's call this the Dirty Dozen:
- What is the material?
- What kind of load have you imposed on the part?
- Is the load repetitive or one time?
- Does the load reverse? Are there major loads from more than one direction that occur at different times?
- What failure mode are you checking for? Broadly speaking you should be checking for ultimate failure, yield failure, fatigue and bucking. Deflection is also something you may find necessary to check for. Impact can also affect the failure mode.
- What temperature range will the part operate at? If high or low temperatures are they temporary or sustained while the load is applied?
- How sure are you about the material properties you used in your FEA? Have you verified them with testing?
- How sure are you about the loads?
- Has the FEA model converged? FEA is an approximation. It is always good practice to refine the mesh in the area of interest and plot the result at a particular point to make sure the model has converged.
- Have you done a quick and dirty hand calculation to make sure the FEA result makes sense?
- Have you correlated your FEA model with test data? At the very least there should be correlation of displacements or deflections.
- Is the geometry in your model the mean or is it the worst case. What is meant by this is, Are the dimensions on your part set to the tolerance limit that would make the geometry the weakest? For example, is the radius of your rib set to the smallest value as allowed by the tolerances shown on the print?