Relationship between probability of breakage to maximum principal stresses in window glass

Currently, the ASTM design methodology to determine the load resistance of annealed window glass incorporates a probability distribution to model glass load resistance. A probability of 8 lites per 1,000 broken at the first occurrence of the design load was selected to match a load resistance consistent with a historical design factor of 2.5. The historical use of a factor relationship leads to the misconception that the design methodology follows an allowable stress procedure. The misconception has led to another common misconception among architects and engineers that a constant maximum principal stress exists, associated with the load resistance for any combination of lite thickness, aspect ratio, and surface area. This paper presents the relationship between the maximum principal stress in glass lites associated with their design loads for a probability of breakage of 8 lites per 1,000. The relationship clearly shows that the maximum principal stress is not constant for a single lite thickness for varying rectangular dimensions much less for all lite geometry combinations. A series of charts illustrates the trends in magnitude and location of the maximum principal stress as a function of lite thickness, aspect ratio, and surface area.