The study conducted by Wei et al。 [46], proposed a criterion quantifying damage based on damage area recorded by IRT。 This criterion is given by the following   equation;AðNiÞ

of endurance? To answer this question, further analysis is    indeedneeded。

3。2。2。 Thermography analysis

The deformation of the structure is usually followed by heat dissipation。 When the material is deformed or damaged, a part of the energy necessary to the starting and the propagation of the damage is irreversibly transformed into heat  [35,38–46]。

Figs。 9 and 10 show the surface temperature contours recorded by the IR camera for an AWJM specimen under load levels of 3, 8 and 11 kN (these loads represent respectively 18%, 47% and 65% of UTS) and for CM specimen under load of 3, 8 and 10 kN (these loads represent respectively 18%, 47% and 59% UTS)。 All the images represented in these figures are obtained at 5000 cycles。 The differ-

where D is the damage function, A(Ni) is the damage area measured at the ith cycle and A(Nf) is the damage area after the number of cy- cles to failure。 The results  presented by the  authors  show  a  correla-

tion between the damage predicted using Eq。 (2) and that measured by the conventional ‘loss of rigidity’ method。 However, no thermal threshold is used in quantifying the evolution of damage。 Extrapo- lating from Wei et al。 [46], this paper proposes damage comparison and fatigue failure prediction using a thermal threshold – the Ther- mographic Damage Criterion (TDC)。 The TDC is an area ratio based on the area of maximum temperature recorded in IRT images (e。g。 Figs。 9 and 10):

At ðDTmaxÞ

ent images recorded by IR camera show that, for loads less than 7 kN,  the   temperature  remains  constant  throughout  the    surface

around the hole (Figs。 9a and 10a)。 For a load of 8 kN, temperature fluctuations increase but remain moderate (Figs。 9b and 10b)。 However, for loads of 10 kN and 11 kN, there is a significant in- crease in temperatures and the fluctuations in the area  reach  100% (Figs。 9c and 10c)。 This area is situated at ±45° from the axis of load, which also represents the axis of the direction of the plies。 However, temperature fluctuations in the CM specimens were sig- nificant for loads less than those machined with abrasive water jet for the same  loading。

Fig。 11 plots the regional temperature and accumulated damage against normalized number of cycles。 A strong and direct correla- tion between the recorded temperature and the damage accumu- lated around the hole is observed。  The  temperature  of  the studied area (region within the rectangle box shown in Figs。 11 and 12) increases with the increase of damage accumulation。 This finding agrees with previous observations [35]。

Fig。 11 depicts the evolution of the maximum temperature and damage before final failure of the specimen。 Two stages for the temperature evolution were distinguished。 In first stage, the varia-

Table 2

2D Statistical roughness parameters of average surface roughness Ra, maximum profile valley depth Rv and skewness Rsk for the two types of   specimens。

Specimen Ra (lm) Rv (lm) Rsk

CM 3。93 ± 0。2 15。44 ± 0。3 0。54

AWJ 3。50 ± 0。2 12。85 ± 0。3 —0。20

where Ai  (DTmax) is the area (pixels) of the maximum  temperature