3.1 Introduction

A Designer has to ensure that the structures and facilities one designs are (i) fit for their purpose (ii) safe and (iii) economical and durable. Thus safety is one of the paramount responsibilities of the designer. However, it is difficult to assess at the design stage how safe a proposed design will actually be – consistent with economy. There is, in fact, a great deal of uncertainty about the many factors, which influence both safety and economy. Firstly, there is a natural variability in the material strengths and secondly it is impossible to predict the loading, which a structure (e.g. a building) may be subjected to on a future occasion. Thus uncertainties affecting the safety of a structure are due to

uncertainty about loading

uncertainty about material strength and

uncertainty about structural dimensions and behaviour.

These uncertainties together make it impossible for a designer to guarantee that a structure will be absolutely safe. All that the designer could ensure is that the risk of failure is extremely small, despite the uncertainties.

An illustration of the statistical meaning of safety is given in Fig. 3.1. Let us consider a structural component (say, a beam) designed to carry a given nominal load. Bending moments (B.M.) produced by characteristic loads are first computed. These are to be compared with the characteristic resistance or strength (R.M.) of the beam. But the characteristic resistance (R.M.) itself is not a fixed quantity, due to variations in material strengths that might occur between nominally same elements. The actual resistance of these elements can be expected to vary as a consequence. The statistical distribution of these member strengths (or resistances) will be as sketched in (a).

Figure 3.1 Statistical Meaning of Safety

Similarly, the variation in the maximum loads and therefore load effects (such as bending moment) which different structural elements (all nominally the same) might encounter in their service life would have a distribution shown in (b). The uncertainty here is both due to variability of the loads applied to the structure, and also due to the variability of the load distribution through the structure. Thus if a particularly weak structural component is subjected to a heavy load which exceeds the strength of the structural component, clearly failure could occur.

Unfortunately it is not practicable to define the probability distributions of loads and strengths, as it will involve hundreds of tests on samples of components. Normal design calculations are made using a single value for each load and for each material property and making appropriate safety factor into the design calculations. The value used is termed as “Characteristic Strength or Resistance” or “ Characteristic Load”. Characteristic resistance of a material (such as Concrete or Steel) is defined as that value of resistance below which not more than a prescribed percentage of test results may be expected to fall. (For example the characteristic yield stress of steel is usually defined as that value of yield stress below which not more than 5% of the test values may be expected to fall). In other words, this strength is expected to be exceeded by 95% of the cases. Similarly, the characteristic load is that value of the load, which has an accepted probability of not being exceeded during the life span of the structure. Characteristic load is therefore that load which will not be exceeded 95% of the time.

3.2 Standardisation

Most structural designs are based on experience. Standardisation of all designs is unlikely within the foreseeable future hence design rules, based on experience, become useful. If a similar design has been built successfully elsewhere, there is no reason why a designer may not consider it prudent to follow aspects of design that have proved successful, and adopt standardised design rules. As the consequences of bad design can be catastrophic, the society expects designers to explain their design decisions. It is therefore advantageous to use methods of design that have proved safe in the past. Standardised design methods can help in comparing alternative designs while minimising the risk of the cheapest design being less safe than the others.

Most Governments attempt to ensure structural safety through regulations and laws. Designers then attempt to achieve maximum economy within the range of designs that the regulations allow. Frequently the professions are allowed to regulate themselves; in these a cases the Regulations or Codes of Practices are evolved by consultation and consensus within the profession.

3.3 Allowable stress design (ASD)

With the development of linear elastic theories in the 19th century the stress-strain behaviour of new materials like wrought iron & mild steel could be accurately represented.

theories enabled indeterminate structures to be analysed and the distribution of bending and shear stresses to be computed correctly. The first attainment of yield stress of steel was generally taken to be the onset of failure. The limitations due to non-linearity and buckling were neglected.

The basic form of calculations took the form of verifying that the stresses caused by the characteristic loads must be less than an “allowable stress”, which was a fraction of the yield stress. Thus the allowable stress may be defined in terms of a “factor of safety” which represented a margin for overload and other unknown factors which could be tolerated by the structure. The allowable stress is thus directly related to yield stress by the following expression:

Allowable stress = Yield stress/ Factor of Safety

In general, each member in a structure is checked for a number of different combinations of loading. The value of factor of safety in most cases is taken to be around 1.67. Many loads vary with time and these should be allowed for. It is unnecessarily severe to consider the effects of all loads acting simultaneously with their full design value, while maintaining the same factor of safety or safety factor. Using the same factor of safety or safety factor when loads act in combination would result in uneconomic designs.

A typical example of a set of load combinations is given below, which accounts for the fact that the dead load, live load and wind load are all unlikely to act on the structure simultaneously at their maximum values:

(Stress due to dead load + live load) < allowable stress (Stress due to dead load + wind load) < allowable stress

(Stress due to dead load + live load + wind) < 1.33 times allowable stress.

In practice there are severe limitations to this approach. These are the consequences of material non-linearity, non-linear behaviour of elements in the post-buckled state and the ability of the steel components to tolerate high theoretical elastic stresses by yielding locally and redistributing the loads. Moreover the elastic theory does not readily allow for redistribution of loads from one member to another in a statically indeterminate structures.

3.4 Limit state design

An improved design philosophy to make allowances for the shortcomings in the “Allowable Stress Design” was developed in the late 1970’s and has been extensively incorporated in design standards and codes formulated in all the developed countries. Although there are many variations between practices adopted in different countries the basic concept is broadly similar. The probability of operating conditions not reaching failure conditions forms the basis of “Limit States Design” adopted in all countries.

“Limit States” are the various conditions in which a structure would be considered to

have failed to fulfil the purpose for which it was built. In general two limit states are considered at the design stage and these are listed in Table 3.1.

Table 3.1: Limit States

Limit State of Strength

Serviceability Limit State

Strength (yield, buckling)


Stability against overturning and sway


Fracture due to fatigue

Fatigue checks (including reparable

Plastic collapse

damage due to fatigue)

Brittle Fracture



“Limit State of Strength” are: loss of equilibrium of the structure and loss of stability of the structure. “Serviceability Limit State” refers to the limits on acceptable performance of the structure.

Not all these limits can be covered by structural calculations. For example, corrosion is covered by specifying forms of protection (like painting) and brittle fracture is covered by material specifications, which ensure that steel is sufficiently ductile.

3.5 Partial safety factor

The major innovation in the new codes is the introduction of the partial safety factor format.

A typical format is described below:

In general calculations take the form of verifying that


where S* is the calculated factored load effect on the element (like bending moment, shear force etc) and R* is the calculated factored resistance of the element being checked, and is a function of the nominal value of the material yield strength. S* is a function of the combined effects of factored dead, live and wind loads. (Other loads – if applicable, are also considered)

In accordance with the above concepts, the safety format used in Limit State Codes is based on probable maximum load and probable minimum strengths, so that a consistent level of safety is achieved. Thus, the design requirements are expressed as follows:

Sd ≤ Rd

where Sd = Design value of internal forces and moments caused by the design Loads, Fd

Fd = γf* Characteristic Loads.

γf= a load factor which is determined on probabilistic basis Rd = Characteristic Value of Resistance/γm

where γm = a material factor, which is also determined on a ‘probabilistic basis’

It should be noted that γf makes allowance for possible deviation of loads and the reduced possibility of all loads acting together. On the other hand γm allows for uncertainties of element behaviour and possible strength reduction due to manufacturing tolerances and imperfections in the material.

Collapse is not the only possible failure mode. Excessive deflection, excessive vibration, fracture etc. also contribute to Limit States. Fatigue is an important design criterion for bridges, crane girders etc. (These are generally assessed under serviceability Limit States)

Thus the following limit states may be identified for design purposes:

Ultimate Limit State is related to the maximum design load capacity under extreme conditions. The partial load factors are chosen to reflect the probability of extreme conditions, when loads act alone or in combination.

Serviceability Limit State is related to the criteria governing normal use. Un-factored loads are used to check the adequacy of the structure.

Fatigue Limit State is important where distress to the structure by repeated loading is a possibility.

The above limit states are provided in terms of partial factors reflects the severity of the risks.

An illustration of partial safety factors for applied load and materials as suggested in the revised IS: 800 for Limit States of Strength and Limit States of Serviceability are given in Table 3.2 and 3.3 respectively.

3.6 Factors governing the ultimate strength

Stability is generally ensured for the structure as a whole and for each of its elements. This includes overall frame stability against overturning and sway, as given below. The structure as a whole or any part of it are designed to prevent instability due to overturning, uplift or sliding under factored load as given below:

The actions are divided into components aiding instability and components resisting instability.

The permanent and variable actions and their effects causing instability are combined using appropriate load factors as per the Limit States requirements to obtain maximum destabilizing effect.

The permanent actions (loads) and effects contributing to resistance shall be multiplied with a partial safety factor 0.9 and added together with design resistance (after multiplying with appropriate partial safety factor). Variable actions and their effects contributing to resistance are disregarded.

The resistance effect shall be greater than or equal to the destabilizing effect. Combination of imposed and dead loads should be such as to cause most severe effect on overall stability.

3.7 Limit state of serviceability

As stated in IS: 800: 2007, Serviceability Limit State is related to the criteria, governing normal use. Serviceability limit state is limit state beyond which service criteria, specified below, are no longer met:

Deflection Limit

Vibration Limit

Durability Consideration

Fire Resistance

Load factor, γf, of value equal to unity are used for all loads leading to Serviceability Limit States to check the adequacy of the structure under serviceability limit states, unless specified otherwise.

The deflection under serviceability loads of a building or a building component should be such that, they do not impair the strength of the structure or components or cause damage to finishing. Deflections are to be checked for the most adverse but realistic combination of service loads and their arrangement, by elastic analysis, using a load factors as per Table 3.3. Table 3.4 gives recommended limits of deflections for certain structural members and systems.

As per IS: 800, suitable provisions in the design are required to be made for the dynamic effects of live loads, impact loads and vibration due to machinery operating loads. In severe cases possibility of resonance, fatigue or unacceptable vibrations shall be investigated. Unusually flexible structures (generally the height to effective width of lateral load resistance system exceeding 5:1) need to be investigated for lateral vibration under dynamic wind loads.

Durability or Corrosion resistance of a structure is generally, under conditions relevant to their intended life as are listed below:

The environment

The degree of exposure

The shape of the member and the structural detail

The protective measure

Ease of maintenance

Fire resistance of a steel member is a function of its mass, its geometry, the actions to which it is subjected, its structural support condition, fire protection measures adopted and the fire to which it is exposed.

3.8 Classification Of Cross-Sections

Determining the resistance (strength) of structural steel components requires the designer to consider first the cross sectional behaviour and second the overall member behaviour – whether in the elastic or inelastic material range, cross sectional resistance and rotation capacity are limited by the effects of local buckling.

In the IS 800: 2007 code cross sections are placed into four behavioural classes depending upon the material yield strength, the width to- thickness ratios of the individual components (e.g., webs and flanges) within the cross section, and the loading arrangement. The four classes of sections are defined as follows (see also Fig.3.2):

Figure 3.2: Moment-Rotation Behaviour Of The Four Classes Of Cross-Sections As

Defined By IS 800: 2007

Plastic or class 1 Cross sections which can develop plastic hinges and have the rotation capacity required for the failure of the structure by the formation of a plastic mechanism (only these sections are used in plastic analysis and design).

Compact or class 2 Cross sections which can develop their plastic moment resistance, but have inadequate plastic hinge rotation capacity because of local buckling.

Semi-compact or class 3 Cross sections in which the elastically calculated stress in the extreme compression fibre of the steel member, assuming an elastic distribution of stresses,

can reach the yield strength, but local buckling is liable to prevent the development of the plastic moment resistance.

Slender or class 4 Cross sections in which local buckling will occur even before the attainment of yield stress in one or more parts of the cross section. In such cases, the effective sections for design are calculated by deducting the width of the compression plate element in excess of the semi-compact section limit.

It has to be noted that only plastic sections should be used in indeterminate frames forming plastic-collapse mechanisms. In elastic design, semi-compact sections can be used with the understanding that the maximum stress reached will be My. Slender sections also have stiffness problems and are not preferable for hot-rolled structural steelwork. Compact or plastic sections are used for compression members, since they have more stiffness than semi-compact or slender members.

The maximum value of limiting width-thickness ratio of different classifications of sections is given by the code as shown in Table 3.5. When different elements of a cross section fall under different classifications, the most critical one has to be selected to represent the classification of the cross-section. Most of the hot-rolled sections available in the market fall under the category of plastic or compact sections.

Table 3.2: Partial safety factors for loads, γ f for limit states