EXPERIMENTAL STUDY OF DYNAMIC BENDING STIFFNESS OF ACSR DRAKE OVERHEAD CONDUCTOR Frédéric LÉVESQUEa, Sylvain GOUDREAUb, Sébastien LANGLOISa, Frédéric LÉGERONa a Département de génie civil, Université de Sherbrooke, Sherbrooke, Qc, Canada b GREMCA laboratory, Département de génie mécanique, Université Laval, Québec, Qc, Canada
[email protected],
[email protected],
[email protected],
[email protected] Introduction Aeolian vibrations may cause fretting fatigue failure at or near the location of clamped devices (suspension clamps, dampers, spacers, etc.) [1]. The mitigation of these vibrations is thus a necessity to improve transmission lines reliability and life expectation. The most common way to control these vibrations is to install dampers on conductors. Researches are still under progress to develop design tools for dampers. Before even assessing the behaviour of a damper under design on a conductor, the mechanical behaviour of the conductor itself needs to be modelled with sufficient precision. The non linear variation of bending stiffness of conductors is well known. Models have been proposed to represent this complex behaviour, whether on a stick-slip basis [2] or taking into account the tangential compliance of contacts [3]. In any case, a variation of bending stiffness means that slipping or sliding should be observed inside the conductor. We use the term slipping when only a part of the contact area sees relative motion between surfaces, while we use sliding when there is a bulk relative displacement of the bodies in contact. Hence, fretting marks are symptomatic of repeated local bending stiffness variations. Fretting is a mechanical and metallurgical damaging process observed inside contacts between clamped bodies under variable tangential loadings causing small relative displacements at the interface [4]. But yet, experimental validation of these models was only done on static benches [2] or on very short spans [5]. Results from measurements on three vibrating ACSR (Aluminium Conductor Steel Reinforced) Drake specimens are reported here at various amplitudes, modes and tensions. This builds a database on which one could rely to validate vibrating conductor models (see companion paper by Langlois et al. [6] for example).
EXPERIMENTAL PROGRAM The deformed shape of three vibrating ACSR Drake specimens was measured on a resonance fatigue test bench in the GREMCA laboratory (Université Laval, Québec, Canada). The test bench was previously described in [1] and [7]. The properties of the conductor are given in table 1. A square-faced bushing was used, to avoid the complex end effects observed with commercial suspension clamps. The conductor active length of 5.83m was excited at one of its resonance frequencies by an electromagnetic shaker placed at 0.33m from the end opposed to the bushing. The control parameters were the frequency and the peak-to-peak bending amplitude Yb measured at 89mm from the last point of contact (LPC) between the conductor and the bushing. The frequency multiplying the anti-node amplitude, fymax, is also a widely used vibration parameter, and is thus computed afterwards. The link between the shaker and the conductor was monitored with a load cell and an accelerometer. Bending amplitudes Yb ranged from 0.10 to 0.60mm by increments of 0.10mm. Vibration amplitudes Y were measured at 44.5mm (Y45), 89mm (Yb), 178mm (Y178) and 267mm (Y267) from the exit of the bushing and the maximum displacement ymax at the mid-length anti-node. On the first tested specimen, amplitude was also measured at 133.5mm (Y134). Specimens vibrated at their 3rd, 4th and 5th modes. Static tensions H of 15%, 25% and 35% of conductor Rated Tensile Strength (RTS) were imposed. Vibration nodes were also localized. On the third specimen, the static tension was kept constant at 25%RTS and the effect of the square-faced bushing torque was assessed. Increasing torques of 20, 35 and 50 ft.lb were applied.
In a companion paper by Langlois et al. [6], two variable bending stiffness theoretical models are implemented into a time history finite element model. Numerical results are then compared to the experimental ones presented here. The damping of the overall system (dissipation in the conductor and the test bench) was deducted from the hysteresis loop between the excitation force and the displacement at the point of application of the force. When the signal of the accelerometer placed on the link between the shaker and the conductor has a pure sinusoidal shape, it is easy to integrate the displacement. As a check, the behaviour of the conductor was also monitored while the excitation was cut. The logarithmic decrement method is then used from these data to assess damping ratios. With that damping, a theoretical model should predict the measured driving force amplitude for a given frequency and ymax.
EXPERIMENTAL RESULTS The deformed shape of the first tested specimen (Drake-1) is presented in figures 1 to 3, depending on static tension. Data at the third mode are presented in table 2. Each vibration mode has its own curve. Near the bushing, there is little variation of amplitude depending on the mode (
