Are there any accurate ways to model the tensile properties of composites for a random fibre alignment condition without measuring the ratio of randomness? 

Why does the load-displacement curve falls even before the ultimate strength is reached as shown in the figure? 

True stress is generally represented by  while the engineering stress is represented by . If the area remains constant upto the ultimate strength, why is it that both the curve deviates at the yield strength itself?