Definition of Structural Safety for Design Codes

We have seen in Chapter 1 that structural safety is a very broad concept, encompassing far more than the narrow definition we will adopt here. In the 'big picture' safety includes quality control, avoidance of human error, robustness against unexpected catastrophe and many more attributes. The definition in design codes has little to do with all this. The definition we adopt is as follows:

'A structure is safe if during the expected life of the structure the chance of exceeding a limit state set by the design code is acceptably small.' Among all the major calamities that could beset a structure, this appears to be about a negligible complaint. However, the design codes take care of a lot of detailed requirements which, if ignored, would cause a multitude of problems during the life of the structure.

The main concern is that at the time of design we do not know with certainty the properties of the structure nor the lifetime magnitudes of the loads. We do know these vital quantities vary from those we assume in design. Somehow this uncertainty must be accounted for. Traditionally the codes have provided 'safety factors' for this.

The following sections will discuss various ways in which safety factors are developed and used. It should be reiterated that safety factors are no protection against gross human error, lack of judgement, inexperience, greed, carelessness and other unfortunate happenings which cause the predominant share of the known structural failures.

Allowable Stress Design

Conventional wisdom has it that the greater the ignorance about an event, the larger the factor of safety should be. In the beginning of the scientific building process engineers assessed, as best they could, the strength of the structure. They then divided this strength by 6, 4 or 2 or any other judgemental and experiential number to obtain a safe load. The selection of the factor of safety depended on the confidence of the estimate of the strength. In the second half of the nineteenth century the theory of elasticity began to take hold on the practice of structural design engineers.

This theory assumes linearity between loads and the resulting forces or deformation magnitudes in the structure, as well as between the forces and the resulting stresses. The design philosophy which evolved out of the application of elasticity theory is called 'allowable stress design' (ASD). Simply stated, ASD computes the stresses a by linear theory for the maximum loads that can be expected during the life-span of the structure, and it compared these stresses to 'allowable stresses' a,,, which are a fraction of limiting stresses slim.

These are defined as the stress levels where linear elastic theory ceases to apply, that is where the material yields or the structure becomes unstable. The design criterion is defined as

The abbreviation FS is the 'factor of safety'. The magnitude of this factor, of course, is one of the crucial issues in design codes ; the other is the definition of a,,,. How is such a factor developed? It essentially evolved historically, from high values when a technology is just starting to gradually lower values until a lower ceiling, dictated by common sense and by successful and unsuccessful experience, is reached.

The evolution of the allowable stress for mild steel is illustrated in Table 3.3. For essentially the same material, made by practically the same process, the basic allowable stress in the United States increased by about 60 per cent in 70 years of evolution. Improved quality control in the mills, as well as increased confidence in the design process due to accumulated experience, led to this increase in the allowable stress.