CEMENT and CONCRETE RESEARCH. Vol. 6, pp. 773-782, 1976. Printed in the United States.

Pergamon Press, Inc.

ANALYSIS OF CRACK FORMATION AND CRACK GROWTH IN CONCRETE BY MEANS OF FRACTURE MECHANICS AND FINITE ELEMENTS

A H i l l e r b o r g , M Mod~er and P-E Petersson Division of Building Materials Lund I n s t i t u t e of Technology, Lund, Sweden

(Communicated by Z. P. Bazant) (Received August 24, 1976)

ABSTRACT A method is presented in which fracture mechanics is introduced into f i n i t e element analysis by means of a model where stresses are assumed to act across a crack as long as i t is narrowly opened. This assumption may be regarded as a way of expressing the energy absorption G c in the energy balance approach, but i t is also in agreement with results of tension tests. As a demonstration the method has been applied to the bending of an unreinforced beam, which has led to an explanation of the difference between bending strength and t e n s i l e strength, and of the variation in bending strength with beam depth.

Une m6thode est present,e, par laquelle la m6chanique des ruptures est i n t r o d u i t e dans l'analyse des ~l~ments f i n i s ~ l ' a i d e d'un module, oO les contraintes sont suppos6es d'op6rer sur les c6t6s d'une fissure tant que cette f i s s u r e est ~ t r o i t e . Cette hypoth~se peut etre consid~r~e comme un moyen d'exprimer l ' a b sorption G c d'6nergie en usant l'approche de l ' ~ q u i l i b r e d'6nergie. Cette hypoth6se est aussi j u s t i f i 6 e par les r 6 s u l t a t s des essais de tension. Pour en prouver la v a l i d i t Y , cette m~thode a ~t~ appliqu~e au fl6chissement d'une poutre non arm~e et f o u r n i t une explication de la diff6rence entre la r~sistance au moment de flexion et la r~sistance ~ l ' e f f o r t de tension, ainsi que de la variation de la r~sistance au moment de flexion en fonction de la profondeur de la poutre.

773

774 A. H i l l e r b o r g , M. Mod4er, P-E. Petersson Importance of cracks and crack growth

Vol. 6, No. 6

Crack formation and crack growth play an important part in the performance of unreinforced and reinforced concrete. Examples of t h i s are crack spacing and crack width in bending shear chracks and t h e i r e f f e c t on shear capacity cracking moment of reinforced and unreinforced beams microcracks in compression and compression f a i l u r e . A rational design in these cases ought to be based on r e a l i s t i c t h e o r e t i c a l models, which take crack formation and crack propagation i n t o account. So f a r no such models have been a v a i l a b l e . Consequently the design methods have had to be based on empirical research, supported by s i m p l i f i e d models. Recent advances w i t h i n f r a c t u r e mechanics and f i n i t e element methods (FEM) have now given us a p o s s i b i l i t y of analysing crack growth. Fracture mechanics gives the fundamental rules f o r crack propagation and FEM makes i t possible to apply these rules to compl i c a t e d cases. The cases we wish to analyse are rather complicated, as they involve diverse phenomena, such as formation and propagation of cracks two or more p a r a l l e l bent shear cracks shrinkage s t r a i n s i n t e r a c t i o n between concrete and reinforcement i n t e r a c t i o n between cement matrix and aggregate. I t is therefore necessary to use FEM and also to t r y and find a method which s i m p l i f i e s the analysis as much as possible. Proposed approach There are many methods to choose from f r a c t u r e mechanics, e.g. the stress i n t e n s i t y f a c t o r approach the energy balance approach the " s t r i p - y i e l d " model according to Dugdale the cohesive force model according to Barenblatt. The d i f f e r e n t methcds are known to give coherent r e s u l t s . In the stress i n t e n s i t y f a c t o r approach the stresses near the crack t i p are studied. These stresses t h e o r e t i c a l l y a ag_j~roach i n f i n i t y at the crack t i p according to the expression o = K/V2~r, where r is the distance from the crack t i p and K is a c o e f f i c i e n t , the stress i n t e n s i t y f a c t o r , depending on the load, the crack dimensions, etc. When K reaches a c r i t i c a l value Kc, the crack propagates. The stress i n t e n s i t y f a c t o r approach has been used a great deal in FEM analysis. The d i r e c t method requires a FEM mesh with very small elements close to the crack t i p , which l i m i t s i t s a p p l i c a b i l i t y to cracks

Vol. 6, No. 6 FRACTURE MECHANICS, ANALYSIS, CONCRETE complicated problems. I n d i r e c t and special methods permit the use of greater elements. The methods cannot explain the formation of cracks, only the propagation. In the energy balance approach i t is assumed t h a t a certain amount of energy Gc is absorbed by the formation of a u n i t area of crack surface. When a crack propagates a certain amount of stored energy is released. The crack propagates when the released energy is equal to or greater than the absorbed energy. FEM has been used to determine the energy release rate in the energy balance approach, see e.g. / 5 / . This enables the use of a FEM mesh with rather large elements. The formation of cracks cannot be explained. In the Dugdale model i t is assumed that there is a p l a s t i c zone near the crack t i p according to Fig. I . Within the p l a s t i c zone a stress equal to the y i e l d strength ~y acts across the crack. The Barenblatt model is s i m i l a r to the Dugdale model, but the stress is assumed to vary with the deformation. I t does not seem to have been used in f i n i t e element analysis. The basic idea of the model we propose is demonstrated in Fig. 2. I t is in some respects s i m i l a r to the Barenblatt model. The model is described only for mode I (the opening mode), but i t may also be applied to modes I I and I I I .
elastic crock
|

775

crctck length

real crack

~piastic zone]

i
,
m

L
I

Fig. 1 The Dugdale model f o r crack t i p p l a s t i c i t y

Fig. 2

Proposed model

The crack is assumed to propagate when the stress at the crack t i p reaches the t e n s i l e strength f t " When the crack opens the stress is not assumed to f a l l to zero at once, but to decrease w i t h increasing crack width w, for example according to Fig. 3. At the crack width wI the stress has f a l l e n to zero. For t h a t part of the crack where W<Wl, the "crack" in r e a l i t y corresponds to a microcraced zone with some remaining ligaments for stress t r a n s f e r . As there is a stress to be overcome in opening the crack, energy is absorbed. The amount of energy absorbed per u n i t crack area in widening the crack from zero to or beyond wI is

and c o r r e s p o n d s to t he area between the curve and the c o o r d i n a t e a x i s ' in Fig. 3.

776 A. H i l l e r b o r g , M. Modeer, P-E. Petersson

Cr

Vol, 6, No. 6

Fig. 3

ft

Assumed v a r i a t i o n of stress o with crack width w, general case

We now choose the curve in Fig. 3 so that

0 w!
I W

71dw = G
0

(I)

which means that the energy absorbed per newformed u n i t crack area is the same as in the energy balance approach. The model of Fig. 2 may thus be looked upon as a way of expressing the energy balance approach. At the same time the assumption of Fig. 2 may be looked upon as a r e a l i t y . Stresses may be present in a microcracked zone as long as the corresponding displacement is small. This has been c l e a r l y demonstrated in tension t e s t s , using a very r i g i d t e s t i n g equipment, e . g . , by Evans and Marathe / 4 / ; cf. Fig. 5. By the a p p l i c a t i o n of the proposed model the curve ~(w) may be chosen in d i f f e r e n t ways, e.g. according to Figs. 4a, b or c, which a l l show simple mathematical r e l a t i o n s . For t y p i c a l y i e l d i n g materia l s , l i k e mild s t e e l , Fig. 4a seems to be the best choice. I t corresponds e x a c t l y to the Dugdale model with f t = ~y.and ~I = COD at i n i t i a t i o n of crack growth. The d i s c o n t i n u i t y may glve r l s e to some problems by the a p p l i c a t i o n in FEM, but they are not serious.

fl

Fig. 4

L w I w W W I

Examples of poss i b l e assumptions of v a r i a t i o n of stress ~ with crack width w in practical applications

b)

C)

For concrete i t seems t h a t Fig. 4c is the best choice as i t corresponds reasonably well with tension t e s t r e s u l t s / 4 / , cf. Fig. 5. I t is also simple, continuous and s u i t a b l e f o r FEM analysis. For our purpose we have therefore chosen Fig. 4c. We then obtain ~dw = ftwl/2
0

or from ( I ) , wI =

2Gc/f t

(2)

For ordinary concrete Gc/f t seems to be of the order 0.005 -

Vol. 6, No. 6 FP~ACTURE MECHANICS, ANALYSIS, CONCRETE

777

O.Ol mm, cf. / I / , and thus wI of the order 0.01 - 0.02 mm. In the a p p l i c a t i o n we f u r t h e r assume that the concrete is l i n e a r - e l a s t i c u n t i l f t is reached. Fig. 5 shows a comparison between our assumptions with Gc/f t = 0.01 mm, ~ / f t = I0 000 and a tension t e s t from / 4 / with a gage length of 1 (25 mm). This corresponds to a t h e o r e t i c a l average elongation over thergage length when ~ reaches O, i . e . w = Wl, of 2.0.01/25 = 800-I0 - . The assumptions seem to agree reasonably with the t e s t r e s u l t . A lower value of E/f t would have improved the agreement, but from the point of view of the energy balance approach the E-value corresponding to unloading is most important and t h i s j u s t i f i e s the choice E/f t = I0 000.

ft

Fig. 5

A test result from a t e n s i l e t e s t according to Evans & Marathe / 4 / , compared to a corresponding assumed r e l a t i o n by the analysis
or , , , ~ .

200

ZOO

600

BOO

Strain x 106 (1" gage length)

A special feature of the proposed method is t h a t i t explains not only the growth of e x i s t i n g cracks, but also the formation of new cracks, as i t is assumed t h a t cracks s t a r t forming when the t e n s i l e stress reaches f t , i . e . , t h e same c r i t e r i o n is used f o r formation and propagation of cracks. The analysis can be performed with a rather coarse mesh, as in the example below, because there are no stress s i n g u l a r i t i e s and the amount of absorbed energy is not very s e n s i t i v e to the mesh size. The p o s s i b i l i t y of using a coarse mesh means t h a t rather compl i c a t e d problems can also be treated without using too many elements. A p p l i c a t i o n to an unreinforced beam in bending In order to study the a p p l i c a b i l i t y ing case has been analysed / 2 / . of the method the f o l l o w -

An unreinforced concrete beam w i t h a constant rectangular crosssection is loaded by a pure bending moment M according to Fig. 6. When the bending moment reaches a value M o the t e n s i l e stress in the bottom f i b r e reaches f t - As we assume t h a t the concrete cannot take higher t e n s i l e stresses than f t , cracks w i l l form and s t a r t opening when M is increased above M o. We w i l l now study how these cracks grow when the bending moment increases. In order to s i m p l i f y the c a l c u l a t i o n s we assume t h a t only one crack opens, and t h a t t h i s happens at the section of symmetry. The f i n i t e element mesh used f o r the c a l c u l a t i o n is shown in Fig. 6. The bending moment M is applied as a couple of forces at the

778 A. H i l l e r b o r g , M. Mod#er, P-E. Petersson

Vol. 6, No. 6

~
1,4 NIt

4J 42 41 40 J9 3'8 J7 .16
35 33 32 3t

Fig. 6

Bent rectangul a r beam and a corresponding FEM representation

l e f t end of the beam. The crack i s assumed to open at the s e c t i o n to the r i g h t , which is the s e c t i o n o f symmetry. Mois the moment which gives ~31 = f t , where ~31 i s the s t r e s s at p o i n t 31. Mo would be the f a i l u r e moment i f the material were e l a s t i c and p e r f e c t l y b r i t t l e . When M is raised above M o the crack s t a r t s opening at p o i n t 31. At t h a t p o i n t we i n t r o d u c e a force corresponding to the r e l a t i o n between s t r e s s ~ and crack width w according to Fig. 4c. With t h i s new f i n i t e element system we can c a l c u l a t e the s t r e s s at p o i n t 32 and we can determine t h a t value M = MI , which gives a s t r e s s 032 : f t " We can now i n t r o d u c e another force at p o i n t 32 and c a l c u l a t e a moment M = M 2, g i v i n g 033 = f t etc. By proceeding in the same way we get a r e l a t i o n between crack depth and applied moment according to Fig. 7.

Mo

Fig. 7

Calculated bending moment M versus crack depth

0 i i

crack depth node point

When the crack grows the corresponding bending moment reaches a maximum value Mmax whereupon i t s t a r t s decreasing. As the maximum value is reached the s t r u c t u r e becomes unstable i f M i s kept c o n s t a n t , and i t f a i l s suddenly as the crack propagates. The r e l a t i o n Mmax/M o i s the same as the r e l a t i o n between bending s t r e n g t h and tensTTe s t r e n g t h , as M o i s the moment which makes the maximum bending s t r e s s in the uncracked section equal to the tensile strength.

Vol. 6, No. 6 FRACTURE MECHANICS, ANALYSIS, CONCRETE I t can be shown t h a t the behaviour of the beam depends on the parameter H/I , where H is the beam depth and 1 is a c r i t i c a l length, defined by c c I c : EG/f~ As the r e a l t i o n EGc = K2
C

779

(3)

holds for plain stress and approximately for plain s t r a i n , we may also write 2 I c = (Kc/ft) (4)

Fig. 8 shows the r e s u l t s of the above analysis as well as of an analysis where shrinkage s t r a i n s ~s according to Fig. 9 have been taken i n t o account.
Bendin~
Tensile strength strength

Fig. 8

Theoretical v a r i a t i o n of r a t i o between bending and t e n s i l e strength with beam depth H and I c = ( K c / f t ) ~ = :

2,0 !

EGc/f~

1.5 .~oO

J
t

1.0 ~,E ~ T .a~

0.5

~....~

c~ j
,, .....<
r L 2 3 4

Fig. 9

Assumed d i s t r i bution of s h r i n k age s t r a i n s

H/I c

I t must be remembered t h a t the r e s u l t s in Fig. 8 correspond to a simple FEM model where only one crack is assumed to open, i n dependent of the stresses in the other parts of the beam. A more r e a l i s t i c model with cracks opening in a l l places where f t is exceeded w i l l give somewhat d i f f e r e n t r e s u l t s with higher values of Mmax/Mo, e s p e c i a l l y where shrinkage s t r a i n s are present.

780 A. H i l l e r b o r g , M. Mod~er, P-E. Petersson

Vol. 6, No. 6

Relative
strenq th

bending

1,0

e-

: = ~ ,

Fig. I0
0.5 -

Test r e s u l t s of bending s t r e n g t h versus beam depth, summarized by Meyer / 3 / , compared to t h e o r e t i c a l curve f o r 1 = I00 nTn
C

100
1931 1932

200

300
768 2/, 111 71 120 ~ Vat~@s Va{ues V~Lue$ VattJeS VaLueJ VQI~es

500

H mm

Reag4t,WiHIs. I KtLlermomn,

= W h r i g h t . C~rwood IO~Z O Nie{sen 195,~

Lindner. Spra~uelg$5 WaLker ~ [ O l m Meyer 1957 1963

~8 V a L u e s

o o ~'

wOt dried O.O0"J'~ * H 2 P,~urs drded O,OtS. H 2 h ~ r s gs%

Relative stren( th

bending

confidence

ktt4r~l

1,5

,.o

I I i

Fig. I I

o,

Test r e s u l t s o f bending s t r e n g t h versus beam depth f o r one q u a l i t y o f concrete, tested wet and d r i e d in 45% RH, compared to t h e o r e t i c a l curves f o r 1 = I00 mm
C

,
o

! i

!
2OO 300

i"

II-

loo

H mm

Vol. 6, No. 6 FRACTURE MECHANICS, ANALYSIS, CONCRETE

781

Fig. I0 shows a comparison between theoretical values according to Fig. 8 and test results summarized by Mayer / 3 / . The theoretical curve is shown for I c = EGc/f ~ : I0 000"0.01 = I00 mm, corresponding to the values used in Fig. 5. I t has been assumed that there is no shrinkage. Fig. 11 shows a comparison between theoretical values according to Fig. 8 and our own test results. Regarding the influence of shrinkage i t must be noticed that the t e s t specimen had a square cross-sect i o n , drying in all d i r e c t i o n s , whereas the theoretical curve is valid for a specimen drying only upwards and downwards, and that creep was not taken into account in the calculations. In spite of i t s s i m p l i f i c a t i o n , the model seems to be able to explain the test results. Conclusion The proposed method of combining fracture mechanics and f i n i t e element analysis seems to y i e l d r e a l i s t i c results regarding crack f o r mation and propagation as well as regarding f a i l u r e even i f a coarse element mesh is used. This opens up the p o s s i b i l i t y of studying complicated problems with a limited amount of computer work. References /I/ Welch, G B and Haisman, B "Fracture toughness measurements of concrete," University of New South Wales, Sydney, A u s t r a l i a , Uniciv. Report No R 42, January 1969. Petersson, P-E and Mod~er, M "Model based on fracture mechanics for the calculation of crack propagation in concrete," (in Swedish with English summary), Division of Building Technology, Lund I n s t i t u t e of Technology, Sweden, Report No 70, 1976. Mayer, H "Die Berechnung von Durchbiegung von Stahlbetongbaut e i l e n , " Deutscher Ausschuss fur Stahlbetong, H 194, W Ernst & Sohn, Berlin, 1967. Evans, P H and Marathe, M S "Microcracking and s t r e s s - s t r a i n curves for concrete in tension," Materials and Structures (RILEM), No I , pp 61 - 64, 1968. for estimating the cracking moment of reinforced concrete beams," ACl Journal No 7, July 1975, pp 356 - 360.

/2/

/3/

/4/

151 Salah EI-Din, A S and EI-Adawy Nassef, M M "A modified approach