Failure analysis of a lifting platform for tree pruning

J. Toribio a,*, V. Kharin a, F.J. Ayaso a, B. González a, J.C. Matos b, D. Vergara a, M. Lorenzo c

a Department of Materials Engineering, University of Salamanca, E.P.S., Campus Viriato, Avda. Requejo 33, 49022 Zamora, Spain

b Department of Computing Engineering, University of Salamanca, E.P.S., Campus Viriato, Avda. Requejo 33, 49022 Zamora, Spain

c Department of Mechanical Engineering, University of Salamanca, E.T.S., Ingeniería Industrial, Avda. Fernando Ballesteros 2, 37700 Béjar, Salamanca, Spain

abstract

The paper presents an analysis of the failure in service of a lifting platform used for tree pruning. Different fracture mechanics techniques were used to reveal the causes of failure such as analysis of the fracture surface, mechanical and microstructural characterization of the material and fracture mechanics tests to determine the critical value of the stress inten- sity factor and to characterize the subcritical crack growth under fatigue. Results yield do conclude that the platform failure cause was the subcritical crack growth in the welded joint where the welding was discontinuous, thereby producing a stress concentration effect similar to that of a crack.

Keywords: Failure analysis Lifting platform Weldment

1. Introduction

This paper presents the analysis of a failure in service of a lifting platform used for tree-pruning. Basically, the mechanical system function is to lift a worker and his cutting tools to a height of up to about 10 m. The mechanism (Fig. 1) has two degrees of freedom provided by two linear actuators that transmit independently  the  movements to  the  basket  where  the worker is placed.

The kinematic chain of the mechanism arm is formed by two linked articulated quadrilaterals I and II (Fig. 1) formed by respective longitudinal hollow bars. The lower one, quadrilateral I, consists of two parallel articulated quadrilaterals Ia and Ib, which move identically.

Lifting platform failure occurred during tree-pruning when the mechanism arm was placed at working height (10 m). Its fracture produced worker fall down causing serious injuries to him. The post mortem analysis revealed the fracture of one of the longitudinal beams of the quadrilateral Ia at its junction with the common bar shared by the quadrilaterals Ia, Ib and the quadrilateral II, as shown in Fig. 1.

During its life in service the lifting platform was subjected to cycles of lifting up and down, so it can be considered that the mechanism links suffered cyclic loading. Under these conditions, welded joints are particularly sensible to failure because the welding process affects the steel microstructure, causing a decrease of its mechanical properties in the welding zone [1–4]. Then, crack initiation and its growth processes can advance at lower load levels [5,6]. This has caused catastrophic failures in a great number of cases [7]. For this reason the applicable standard establishes very strict requirements for welded joints in structures [8].

Failure analysis presented in this paper was performed using various fracture mechanics techniques [9,10]. Firstly, an exhaustive analysis of fracture surfaces was performed, paying special attention to the neighborhood of welds. Macroscopic inspection of the fracture zone and detailed analysis of fracture surface by scanning electronic microscope (SEM) were

* Corresponding author. Tel.: +34 980 54 50 00; fax: 34 980 54 50 02.

E-mail address: toribio@usal.es (J. Toribio).

1350-6307/$ - see front matter © 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.engfailanal.2009.08.017

Fig. 1. Scheme of the lifting platform indicating the place of fracture.

carried out to elucidate the process. Then, an extensive experimental study was performed to determine the microstructure and mechanical properties of the material. Finally, fracture mechanics tests were performed to obtain the critical stress intensity factor (SIF) KC and the crack growth parameters under fatigue loading.

2. Material characterization

The first point was to determine the material microstructure. The results of the metallographic analysis by means of an attack by dissolution of Nital in ethanol to 4% revealed that the material was almost completely a ferritic steel, Fig. 2.

Next, mechanical characterization of the material was performed. Hardness and micro-hardness, respectively, of 81 HRB and 164 HV were obtained. These values are typical for low carbon steels. Then, tensile tests under displacement control with a test speed of 2 mm/min were performed, using specimens cut from the broken platform parts, so that they had      300 mm length, 21–28 mm width, and 4 mm thickness, this latter being imposed by the bar thickness. The resulting stress–strain curves for four tested specimens are shown in Fig. 3. They render the following Young modulus E = 214 GPa,

tensile yield strength rY = 388 MPa, ultimate tensile strength rR = 493 MPa, and strain at maximum stress eR = 13%.

Fracture mechanics tests in mode I (opening) were carried out with four specimens of rectangular strip-like shape with a single edge crack. The critical value of SIF for crack growth KC = 81 MPa m1/2 was obtained. As far as the thickness of these specimens (4 mm, according to the platform parts thickness) was not enough to ensure the conditions of plain strain, this critical SIF was not the valid fracture toughness of material.

Finally fatigue tests were performed using the strip-like specimens under load control. The results for the cyclic crack growth rate da/dN vs. SIF range DK were fitted to a Paris equation da/dN = C(DK)m with the parameters: C = 1.25    10—10     y m = 4.24 using mm/cycle and MPa m1/2 units, respectively, for the crack growth rate per cycle and SIF.

Fig. 2. Microstructure of the virgin material (far from the welded joint).

600

500

400

300

200

100

0

0 5 10 15 20

e (%)

Fig. 3. Tensile testing data of the material: experimental stress–strain curves.

3. Failure description

Visual inspection of the fractured platform revealed that hollow bars (quadrilaterals Ia and Ib in  Fig. 1) were used  to place  the hydraulic tubes  and electrical cables that provided motion to the linear actuators and control the mechanism,  as  shown         in Fig.  4a. This circumstance was used to distinguish the bars.  So, the bar where electric cables were  placed was  denominated  AC (Fig. 4b) and the bar where hydraulic tubes  were  located was  denominated HP (Fig.  4c).  The quadrilateral I  bars present  the inner  and outer sides  in  relation to their mutual disposition  in the  platform. In order to identify these inner and outer   sides of the bars, additional identifiers were used: ‘‘i” for inner sides (ACi and HPi) and ‘‘o” for outer sides (ACo and HPo), respectively. In addition, it must be taken into account that the bars are hollow, so that a last identifier, an apostrophe (0 ), was  used  to  distinguish  the  inner  surfaces  of  the  hollow  bars  (HPo0 ,  HPi0 ,  ACo0   and  ACi0 )  from  the  outer  ones  (HPo,  HPi, ACo and ACi), see Fig. 5.

It was observed that the fracture zones presented common characteristics in both bars as far as their fractures occurred similarly one to another in the bar-bearing zone (Fig. 6). The assembly process of these unions may provide a better

Fig. 4. Fractured bars of the lifting platform: (a) general view; (b) zoom view of the AC bar; (c) zoom view of the HP bar.

Fig. 5. Details of the fracture appearance and the adopted notations.

Fig. 6. Fracture zones in each bar: (a) HP bar; (b) AC bar.

Fig. 7. View of the complete welding beads at outer surface for the bar-bearing joint: (a) HPo view; (b) ACo view.

understanding of failure. This process usually has two steps. Firstly, a passing hole is made in the bar where the bearing will be placed, with a diameter somewhat bigger than that of the bearing itself. Probably, this hole was made by oxiacethilenic cutting. Later, the final step is performed: to place the bearing and to fix it to the bar by means of a bead welding process by voltaic arc.

From the visual inspection of the outer surfaces of the bars (HPo, HPi, ACo and ACi) it appears that the welding bead (from now on referred to simply as ‘‘bead”) extends completely along the circumferential joint of bar and bearing. On the opposite side  the  visual  inspection  of  inner  surfaces  of  the  bars  (HPo0 ,  HPi0 ,  ACo0   and  ACi0 )  reveals  that  the  bead  is  not  completely extended along a circumference of the joint bar-bearing as it can be seen in Figs. 7 and 8. Results of the visual inspection      in the fractured bars are sketched in Fig. 9, compared with the initial state of the unions before fracture.

Fig. 8. View of the deficient welding beads at the inner surfaces for the bar-bearing joint: (a) HP bar; (b) AC bar.

In particular, it was observed that the bead of bar-bearing joint in the interior of the HP bar extended along an arc of about 90°, so that this union present an arc of 270° without any bound between bar and bearing. In the AC bar the inner beads were not complete as well, although the beads extended along a larger arc of about 180°.

Visual analysis of the fracture surfaces confirmed that fracture initiated at the union of a bead with base material (i.e., with the bars). Then the crack grew by fatigue until it attained the critical length that prompted final fracture. The crack ini- tiation by fatigue is highlighted in Fig. 10, with a red paint, where the separation of the bead and the material base can be clearly appreciated. The same process was found at the bottom side of the picture in Fig. 10a.

A fact should be pointed out: the presence of secondary cracks in the lateral sides in radial direction in relation to the bear- ing axis (Fig. 11). Such lateral quasi-straight cracks seem to be incubated starting from the main circumferencial cracks gen- erated by fatigue at the bar-bearing union. In addition, the afore-said lateral cracks appeared in both faces of the considered bars and grew in a through-the-thickness form. It could be observed that the surface cracks were bigger at the outer sides of the bars than in the inner ones.

It may be assumed that these secondary lateral cracks were produced by fatigue. In addition, taking into account their situation between the main circumferential cracks and the final debonding of the welding region (Fig. 12), it seems to be reasonable to consider them as continuations of the initial circumferential cracks of the bead-bar joint, which deflect from the original propagation direction when the zone of high stress concentration is attained (Figs. 11b and 12). It is observed that the propagation of the main crack by fatigue has a branch in the advance direction when it reaches a micro-notch pro- duced during the extraction of material from the bar to allow the bearing to be placed (probably by oxiacethilenic cutting).

4. Discussion

It appears that the lifting platform fracture started with the initiation of fatigue cracks (the afore-said main cracks) in those locations of the bar-bearing joints where an incomplete welding was performed at the inner sides of the bars, because only the outer welded seams were totally completed, whereas the inner welded seams were discontinuous (only about 50% of the seam was completed). This means a lack of lateral fastening of the bar-bearing joint, and a discontinuity (or slot) acting as an initial crack. Therefore, the bar-bearing union becomes weaker not only because the loss of constrain generated by the incompleteness of the welded seam, but, moreover, due to the notch effect (or crack-like effect) of the slot which produces stress concentration in this zone.

In addition, thermal degradation of the steel was observed in the heat-affected zone (HAZ) near the welded seam. Fig. 13 shows the special microstructure in this zone, very different from that of the heat-unaffected microstructure of the base material (cf. Fig. 2) consisting of ferritic grains.  The  micro-hardness  of  this  HAZ  was  measured,  providing  a  value  of 175 HV, quite higher than micro-hardness in the remote locations, i.e., of the base material. Thus microstructure alteration in the HAZ makes material therein harder and, consequently, more brittle and dangerous than in the rest of the lifting platform.

Apparently, the initial crack grows in two directions: (i) mainly along the circumferential paths in the bead-bar welded joint; (ii) branches of lateral cracks in radial direction (cf. Fig. 12). The fracture surfaces of these lateral cracks (Fig. 14), when analyzed by SEM, showed typical fatigue striations in a direction perpendicular to the crack front advance (Fig. 15). This con- firms that this secondary cracking (leading to catastrophic failure) was driven by cyclic loading.

Following the initial cracking by fatigue, a fracture surface can be found with a tearing appearance, which indicates a quite fast separation of the bead and the bar (base material). This zone has a brighter appearance that the previous fatigue cracking stage, probably due to the shorter time of exposure to the environment of this zone.

Finally, when the circumferential fatigue crack along the welded seam in the bead attains a critical value, unstable crack propagation took place up to the global failure of the lifting platform, such a critical crack spreading approximately along

  BEFORE   AFTER

HPo' tear  material  

welding o

HPi'        

welding i'

welding o'

tear material

welding o

       ACo'

welding o'

tear material

       ACi'

welding i

tear

material

welding i'

Fig. 9. Deficient welding joint scheme before and after (slightly zoomed) fracture.

Fig. 10. Views of the HP bar and its bearing in the initiation zone of the fatigue cracking.

Fig. 11. Fatigue crack in radial direction in HPi bar.

Fig. 12. Fatigue crack in radial direction in HPo bar.

Fig. 13. Altered microstructure of the material in the heat-affected zone of the welding area.

180° of the welding circumference (including both the initial fatigue crack and the tearing zone), as shown in Fig. 16. The appearance of such a  tearing zone resembles the typical ductile fracture associated with low thickness specimens, as in    the case of the bars of the lifting platform.

Fig. 14.  Fatigue  crack appearance.

Fig. 15. Microfractography of the fatigue crack growth.

Tear zone

Final Fracture

Side crack

y

z

FInaittiiaglaFaintiigcuieal

x

Fig. 16. Crack initiation and growth until final fracture.