4. Steel Structure Subjected to Fire and Fatigue (1)

8/22/2019 4. Steel Structure Subjected to Fire and Fatigue (1)

1/12

CORROSION, FIRE PROTECTION AN D FATIGUECONSlDERATlONS &-&EL STRUCTURES

3.5 Summaryof corrosionpreventionmethodsThe mechanism of corrosion and the possible ways of its prevention has been discussedin the foregoing sections. The following are the three broad categories of corrosionprevention methods.I. As mentioned earlier, corrosion does not occur in the absence of water. Corrosionprotection can be achieved by a number of methods (e.g.)(a) Application of coatings to separate th e metal from its environment.(b) Avoiding exposure to moisture and air.(c) Attention to detailing of the structures to encourage rapid drainage of water.I1 Cortosion does not occur in the absence of Oxygen and water. This can beachieved by(a) Deaeration of water(b) De-humidification of the atmosphere(c) Application of certain surface coatings111 Corrosion does not occur if the basic electro-chemical reaction is suppressed(a) The use of corrosion inhibitors would suppress either anodic or cathodicreactions 4 d hence the corrosion is prevented.(b) The other hethod is the application of cathodic protection, which floods thesurface with free electrons and prevents formation of anodes.4.0 STEELSTRUCT~RES UBJECTEDTOFIREIn this section a brief teview of aspects of structural steel work subjected to fire is given.The strength of aU engineering materials reduces as their temperature increases.Steel is no exception. However, a major advantage of steel is that it is incombustible andit can fully recover its strength following a fire, most of the times. Fire represents atransfer of energy from a stable condition to a transient condition as combustion occurs.The common examples of fire that affects structural systems are burning of officefurniture, books, and contents of filing cabinet or other materials. During the fire steelabsorbs a significant amount of thermal energy. After this exposure to fire, steel returnsto a stable condition after cocling to ambient temperature. During this cycle of heatingand cooiing,.individual steel members may become slightly bent or damaged, generallywithout affecting the Stability of the whole structure. From the point of view of economy,a significant number of steel members may be salvaged following a post-fire review of afire affected steel structure. Using the principle " f the member is straight after exposuretofire - he steel is O . K , many steel members could be left undisturbed for the rest oftheir service life. Steel members which have slight distortions may be madedimensionally reusable by simple straightening methods and the member may be put tocontinued use with full expectancy of performance with its specified mechanical 'properties. The members which have become unusable due to excessive deformationmay simply be scrapped. In effect, it is easy to retrofit steel structures after fire.

Version I1


2/12

i ,CORROSION.FIREPROTECTION AND FATIGUECONSIDERATIONSOFSTEEL STRUCTURES1

In the case of concrete exposed to fire, it will start changing its colour to pink at pbout285 '~ nd will turn into deep red at about 590'~. Soon after tbat, concrete would turninto quartz aggregate and spalling would start. The degree of spalling is dependent uponthe rate of temperature rise, moisture content and maximum temrxrature for each tvoe af, z - -aggregate. Hence it seen that concrete exposed to fire beyond :ay 600c,may undergoan irreversible degradation in mechanical strength unlike steel where much of its originalstreagth is regained. The above points underline the advantage of steel in terms ofeconomy even in the case of fire. .4.1 Fire loads and fire rating of steel structuresThe term 'fire load' in a compartment of a structure is the maximum heat that can betheoretically generated by the combustible items and contents of the structure. The fireload could be measured as the weight of the combustible material multiplied by thecalorific value r unit weight. Fire load is conveniently expressed in tenns of the floorYspace as MJlm or Mcallm '. More often it would be expressed in terms of equivalentquantity of wood and expressed as Kg wood 1m2 (1 Kg wood = 18MJ). The commonlyencountered fire loads are presented in Table 3. The values are just an indication of theamount of fire load and the values may change from one environment to the other andalso from country to country.

Table3 Fire load on steel structuresExamples of fire load in various s tructm

Basically they represent the endurance of structural steel elements under standardlaboratory conditions. Fig. 9 represents the performance of protected and unprotectedsteel in a laboratory condition of fire. The rate of heating of the unprotected steel isobviously quite high as compared to the fire-protecred steel. We shall see in thefollowing sections that these two types of fire behaviour of steel structure give rise to twodifferent philosophies of fire design. The time equivalence of fire resistance for steelstructures or the fire rating could be calculated as

Type of steel stnretureSchoolHospitalHotelOfficeDepartmental storeTextile mill show room

T, (Minutes) =CWQ, (1 ) :

Knwood / m?1520253535

>200

Version II

The tire rating of steel structures are expressed in units of time '/z, 1, 2, 3 and 4 hours etc.The specified time neither represents the time duration of the real fire nor the timerequired for the occupants to escape. The time parameters are basically a convenient wayof comparative grading of buildings with respect to fire safety.


3/12

CORROSION.FIRE PROTECTIONAND FATTGUECONSIDERATIONS OF STEELSTRUCTURES

where Qj is the fire load M J / ~ ~hich is dependent on the amount of combustiblematerial, 'W' s the ventilation factor relating to the area and height and width of doorsand windows and 'C is a coefficient related to the thermal properties of the walls, floorsand ceiling. As an illustration, the "W' value for a building with large openings could bechosen as 1.5 and fo r highly insulating materials "@' value could be chosen as 0.09.4.2 Mechanical properties of steel at elevated temperaturesWe need to know about the mechanical properties of steel at elevated temperatures in thecase of fire resistant design of structural steel work. Hence in this section we review theimportant mechanical aspects of steel at elevated temperatures. The variations of thenon-dimensional modulus of elasticity, yield strength and coefficient of thermalexpansion with respect to temperature are shown in fi g . 10. The corresponding equationsare given below. Th e variation of modulus of elastic ity ratio E with respect to thecorresponding value at 20c,with respect to temperature T is given by

Time (Minutes)

Fig.9 Rate of heating of structural steel work- E(T )E = =1.0 + T for O ~ C < T < ~ O O ~ C~ ( 2 0 ' ) ( 2 )

1loo690(1.0 -A )1000 for .600 c < T < 1 0 0 0 ~ ~T - 53.5

The yield stress of steel remains unchanged up to a temperature of about 215 '~ and then-loses its strength gradually. T-? yield stress ratio f (with respect to yield stress at 2 0 ' ~ )vs. temperature T relation is given by

Version II


4/12

CORBOSION, FIREPROTECTION AN D FATIGUECONSIDERATIONS OFSTEEL STRUCTURES

Similarly the coefficient of thermal expansion a. also varies with temperature by a simplerelation

A- -

--Yield stress ratio

I I I I I bTemperature OC

1 Fig.10 Mechanical properties of steel at elevated temperaturesThese equations are very useful when one is interested in the analysis of steel structuressubjected to fire.In the codes of practice for steel structures subjected to fire, streng th curves are generallyprovided for st~ ctu i-81 teel work at elevated temperatures. In these curves the strain atwhich the strength is assessed in an important parameter. For exam ple the BS: 5950 part8 has used 1.5% strain as the strain limit as against 2% for Eurocode 3 PartlO. A lowerstrain of 0.5% may be used for columns or components with brittle fire protectionmaterials.

4.3 Fire resistant steelVersion I1


5/12

C M n Si S P Mo +CrFRS S0.20% S1.50%


6/12

CORROSION,FIREPROTECTION AND FATIGUE CONSIDERATIONS OF STEEL STRUCTURES

required for the member. Th e structural steel work then may be an unprotected one.There are two methods of assessing whether or not a bare steel member requires fireprotection.Th e first is the load ratio method which com pares the 'design tem perature' i.e. maximumtemperature experienced by the member in the required fire resistance time, and the'limiting temperatures', which is the temperature at which the mem ber fails.

High H, / A Value Low H,, / A ValueFig.11 The sectionfactor concept

Fig. 12 Some typical values of H p off ire protected steel sectionsThe limiting temperatures for various structural members are available in the relevantcodes of practice. The load ratio may be defined as:Version I1


7/12

CORROSION,FEZ ROTECrrON AND FATIGUECONSDERATICNSOFSTEEL STRUCTURES

Load applied at the fire limit stateh a d *tio =Load causing the member to fail under normal conditions

If the load ratio is less than 1, then no fire protection is required. In the second method,which is appticable to beams, the moment capacity at the required fire resistance time iscompared with the applied moment. When the moment capacity under fire exceeds theapplied moment, no fire protection is necessary4.5Methods of fireprotectionFire protection methods are basically dependent on the fire load, fire rating and the typeof st~~cturalembers. The commonly used fire protection methods are brieflyenumerated below.Spray protection: The thickness of spray protection depends on the fire rating requiredand size of the job. This is a relatively low cost system and could be applied rapidly.However due to its undulating finish, it is usually preferred in surfaces, which are hiddenfrom the view.Board protection: This is effective but an expensive method. Board protection isgenerally used on columns or exposed beams. In general no preparation of steel isnecessary prior to applying the protection.Inhrrnescenf couting: These coatings expand and form an insulating layer around themember when the fire breaks out. This type of fire protection is useful in visiblesteelwork with moderate fire protection requirements. This method does not increase theoverall dimensionsof the member. Certain thick and expensive intumescent coatings willgive about Zhour fire protection. But these type of coatings require blast cleaned surfaceand a priming coat.Concrete encusetnent: This used to be the traditional fire proofing method but is notemployed in structures built presently. The composite action of the steel and concretecan provide higher load resistance in addition to high fire resistance. However thismethod results in increases dead weight loading compared to a protected steel frame.Moreover, carbonation of concrete aids in encouraging corrosion of steel and thepresence of concreteeffectively hides the steel in distressuntil h s too late.5.0 FATIGUE OF STEEL STRUCTURESA component or structure, which is designed to carry a single monotonically increasingapplication of static load, may fracture and fail if the same load oreven smaller load is,applied cyclically a large number of times. For example a'thin rod bent back and forthbeyond yielding fails after a few cycles of such repeated bending. This is termed as the'fatigue failure'. Examples of s'tructures, prone to fatigue failure, are bridges, cranes,' offshorestructure9and slender towers,,etc.,which are subjected to cyclic loading.The fatigue failure is due to progressive propagation of flaws in steel under cyclicloading. This is Partially enhanced by the stress concentration at the tip of such flaw orcrack. As we can see from Fig. 13, the presence of a hole in a plate or simply the

Version I1 2 - 16


8/12

CORROSION, FlRE PROTECTION AND FATIGUECONSIDERATIONSOF STEEL STRUCMIRES

presence of a notch in the plate has created stress concentrations at the points 'm' and 'n'.The stress at these points could be threeor more times the average applied stress. Thesestress concentrationsmay m u r n the material due to some discontinuities in thematerialitself, These stress concentrations are not serious when a ductile material like steel issubjected to a static load, as the stresses redistribute themselves to other adjacentelements within the structure.

Fig. 13 stress con cent ratio*^ in the presence of notches and holes0"

...................: ... ....................../* ......................................................................

f,= Yield stress "Net section ful yplastified atfailure 1

0,r r rmwr rwr rmr r

Fig. I4 Stress pattern at the point of static failure

Version Il


9/12

CORROSION. FIRE PROTECTION AN D FATIGUE CONSIDERATIONS OF STEEL STRUCTURES

At the time of static failure, the average stress across the entire cross section would be theyield stress as show n in Fig.14. However when the load is repeatedly applied or the loadfluctuates between tension and compression, the p i n t s m, n experience a higher range ofstress reversal than the applied average stress. These fluctuations involving higher stressranges, cause minute crack s at these points, which open up p rogressive ly and spread witheach application of the cyclic load and ultimately lead to rupture.

.................. .................. .................. ........ ....._._..Applied cyclic stress

Crack l ~ & ~ t h%A>33

Number of cycles bFig.15 Crack growth and fatigue failure under cyclic load

Fig.16 Terminology used in fatigue resistant design of structural steel workThe fatigue failure occurs after four different stages, namely:

1. Crack initiation at p i n t s of stress concentration2. Crack growth3. Crack propagation4. Final rupture

Version II


10/12

CORROSION, FIRE PROTECTION AND FATIGUECONSIDERATIONSOF STEELSTRUCTURES

plotted ip the vertical ordinateand minimum stresses as abscissa. The line OA representsalternating cycle (R = -I), line OB represents pulsating cycle (R = 0) and OC the staticload (R = 1). Different curves for different yalues of fatigue life 'N' can be drawnthrough point 'C representing the fatigue strength for various numbers of cycles. Thevertical distance between any point on the 'N' curve and the 45' line OC through theorigin represents the stress range. As discussed earlier, the stress range is the importantparameter in the fatigue resistant design. Higher the stress range a component issubjected to, lower would be its fatigue life and lower the stress range, higher would be

detailing rather than an inadequate design of the member for strength. Let us consider alap joint using fillet weld as shown in Fig. 19 . From the schematic stress diagram it isseen that the fillet weld toe becomes a point of stress concentration. As a result, if thejoint is subjected to cyclic loads, the weld toe experiences a variation of larger stressrange compared to the parent member. Hence, a crack may be initiated at the weld toewhere there is stress concentration. This stress concentration .can be eliminated by using

It becomes very important to avoid any local structpral discontinuities and notches bygood design an d this is the most effective means of increasing fatigue life. Where astructure is subjected to fatigue, it is 'important that welded joints are consideredcarefully. Indeed , weld defects and poor weld details are the major contributors offatigue failures. The fatigue performance of a joint can be enhanced by the use oftechniques such as proper weld geometry, improvements in welding methods and betterweld quality control using non-destructive testing (NDT)methods. The followinggeneral points are important for the design of a welded structure with respect of fatiguestrength: (a) use butt welds instead of fillet welds (b) use double sided welds instead ofsingle sided fillet' welds (c) pay attention to the detailing which may cause stressVersionUI - - i i . -.- +-. : ;, - 2 - 20


11/12

CORROSION,FIREPROTF.CrIONAND FATIGUE CONSIDERATIONS OF STEEL S'TRUCNRES

The development of fatigue crack growth and the various stages mentioned above aresymbolically represent& in Fig. 15. Fatigbe failwe can be defined as the number ofcycles and hence time taken to reach a pre-defined or a threshold failure criterion.Fatigue failures are classified into two categories namely the high cycle and low cyclefatigue failures, depending upon the number of cycles necessary to create rupture. Lowcycle fatigue could be classified,as the failures occurring in &w cycles to a few tens ofthousands of cycles, normally under high stress1 strain ranges. High cycle fatiguerequires about several millions of cycles to initiate a failure. The type of cyclic stressesapplied on structural systems and the terminologies used in fatigue resistant design areillustrated in Fig. 16.5.1 S-N Curves and fatigueresistantdesignThe common form of presentation of fatigue data is by using the S-N curve, where thetotal cyclic stress ( S ) is plotted against the number of cycles to failure (N) ifl logarithmicscale. A typical S-Ncurve is shown ibFig. 17.

" ...... ." ' . . j, ;. . ,.. .......... ...:..:........ .........

Cycles of stressforfailure N )Fig. 17 S-N diagramfor fatigue l$e assessment

It is seen from Fig. 17 that the fatigue Life reduces with respect to increase in stress rangeand at a limiting value of stress, the curve flattens off. The point at which the S-N curveflattens off is called the 'endurance lim it'. To carry out fatigue life predictions, a linearfatigue damage model is used in conjunction with the relevstnt S-N curve. One suchfatigue damage model is that postulated by Wohler as shown in Fig. 17: The relationbetween stress and the num ber of cycles for failure could be written as

where 'A r is the number of cycles to failure, 'C 1s the constant dependant on detailingcategory, 'S' s the applied constant amplitude stress range and 'm' s the slope of the S-r' N curve. For the purpose of design it is more convenient to have the maximum andminimum stresses for a given life as the main parameters. For this reason the m d f i e dI Goodman diagram, as shown in Fig. 18, is mostly used. The maximum stresses are IVersion 11


12/12

1CORROSION.FIRE PROTFXTION AND FATIGUECONSIDERATIONS OF STEEL STRUCTURUj

concentration and (d) in very important details subjected to high cyclic stressFs use anynon-destructive testing (NDT) method to ensure defect free details. From the point ofview of fatigue design, the codes of practice classify various'structural joints and detailsdepending upon their vulnerability to fatigue cracks. For example, IS: 1024classif3es thedetailing in the structural steel work in seven classes viz., A, B, C, D, E, F and Gdepending upon their vulnerability to stress concentrations. A typical detailingclassifiedas 'E' is shown in Fig. 20. This class 'E' applies to members fabricated with fullcruciform butt welds. Similarly, the class 'F' is applicable for members with 'T' typefull penetration butt welds, members c o ~ e c t e d y transverse load - canying filletwelds ana members with stud shear connectGrs in composite secfions. Such a typicaldetailing is shown in Fig. 21. The IS: 1024(1968) provides allowqble stress tables for allthe classifications from A-G for different stress ratios of R = F,,,,,/F- and different life(number of cyclesN).Using these tables the allowablestress for a given life time may belinearly interpolated and the life time for a given allowable stress could belogarithmically interpolated. The accuracy of any fatigue life calculation is highlydependent on a good understanding of the expected loading sequence during the wholelife of a structure. Once a global load pattern has been developed, then a more detailed'inspection of particular area of a structure where the effects of loading may be moreimportant called the 'hot spot stresses' which are basically the areas of stressconcentrations. Fill t Weld-: \./ointoncentrationf stress+)

: .0 1 1: i I---+0-

/Schematic stress diagramFig. 19Stress concentrationat the weld toe

n Class E Stressx refers to this

1Load is transmitted

Y directly th~ough thecen tral plate+1 Fig. 20 Class 'E' defailingaccording to IS: I024 (1968)

4. Steel Structure Subjected to Fire and Fatigue (1)

Documents

Transcript of 4. Steel Structure Subjected to Fire and Fatigue (1)