Influence of the size and depth of a circumferential notch on the impact behavior of streetlights. A passive-safety concept

Depósito de Investigación de la Universidad de Sevilla https://idus.us.es/ This is an Accepted Manuscript of an article published by Elsevier in Engineering Failure Analysis, Vol. 53, on July 2015, available at: https://doi.org/10.1016/j.engfailanal.2015.03.023 Copyright 2015 Elsevier. En idUS Licencia Creative Commons CC BY-NC-ND 1 Influence of the size and depth of a circumferential notch on the impact behavior of streetlights. A passive-safety concept. Francisco Moreno-Pérez, Alberto Barroso(1), José Cañas, Federico París. Elasticity and Strength of Materials Group. School of Engineering. University of Seville. Spain. Camino de los Descubrimientos s/n, 41092 Seville, Spain (1) Corresponding author: [email protected] Abstract: Impact tests on weakened streetlights were carried out using a high energy pendulum. The weakening mechanism consists of a circumferential notch at the base of the streetlight. The objective is twofold: firstly to reduce the maximum acceleration values during an impact which will lessen the potential injuries of the occupants of the vehicle and secondly to help avoid the installation of protective barriers, necessary when the supporting structure has no mechanical fuse, thus helping the failure with low energy absorption values. The machining of the notch is possible in streetlights already installed in the road, due to the fact that the actual values of the structural stiffness and strength of these structures are much higher than those values typically required by the external loadings (mainly wind actions). These preliminary tests have shown a significant decrease in the two parameters under analysis, maximum acceleration and absorbed energy, when using the circumferential notch. These results suggest it would be beneficial to perform a crash test following EN12767 in order to certify the streetlight. Keywords: Impact, passive safety, EN12767, crash tests. 1. Introduction Streetlights, as well as other support structures on the road (e.g. traffic lights) are potential obstacles for a vehicle impact. Understanding their energy absorption behavior is of major importance to determine possible injury levels of the vehicle occupants. In impacts where the incident velocity is low and where the streetlight stops the vehicle, secondary injuries caused by the uncontrolled path of the vehicle after the impact could be avoided. On the contrary, in impacts where the incident velocity is high (typically in intercity roads), a sudden stop at the streetlight could originate high deceleration values giving rise to potential human injuries. In this latter case, passive safety, understood as low energy absorption values at the streetlights, may be considered to be desirable. The residual vehicle speed would be, of course higher, but no potential harm to pedestrians is expected in these intercity roads. It must also be taken into account that in some national regulations, see for example [1], using support structures which do not generate deceleration values over a certain threshold in an impact, means that the associated protective barrier is not required to be installed, with the 2 corresponding cost saving. In order to reduce the vehicle incident speed in an efficient way (high energy absorption capability without peak acceleration values), there are specific geometries and structure configurations available (see for example [2], the classical work [3] or the more recent developments with filled structures [4]). Vehicles can also be designed to absorb energy in an efficient way using composite materials [5]. Nevertheless, streetlights are not specifically designed to act as energy absorbers and that is the reason why, in the aforementioned regulations, a protective barrier has to be installed to protect the vehicle from excessive impact loads. The structural stiffness and strength of most streetlight columns is much higher than those required in the standards dictated by the typical loads definitions (e.g. EN 1991-1-4 “Wind Actions” [6]), giving rise to the presence of dangerous obstacles in the road. One way to reduce the potential injuries to occupants in the event of an impact with a streetlight already installed in the road is to weaken, in some way, the area where the vehicle will make impact. In this study, a circumferential notch was made in the streetlight base. Different radii and depths were used and the impact behavior was studied experimentally. There are specific standards concerning passive safety of support structures (e.g. EN 12767 [7]) in which the impact test requirements are defined and described in detail. Nevertheless, the cost of such tests, requiring standardized vehicles, high-speed recordings and many additional technical specifications, are quite high for performing a preliminary study like this one. Therefore, a high-energy pendulum is used first to assess the impact behavior of streetlights. Although the weight and incident velocity chosen is similar to that used in the EN 12767 test (900 kg and 30 km/h), the pendulum (stiff steel block) does not behave in a similar way to a vehicle (designed to deform and absorb energy). Thus, the results and conclusions must only be taken as an initial assessment of the real impact behavior of the streetlight. With all previous comments in mind, the objective of the present study is to analyze the influence of a notch on the impact behavior of a streetlight. 2. Definition and machining of the circumferential notch on the streetlight base Streetlight columns are weakened by making a circumferential notch along the whole perimeter of their base. This weakening procedure is, of course, directed at streetlights already installed on the road and currently has its patent pending. The streetlight chosen for the present study is a straight column, made of galvanized steel, with a height of six meters and 3 mm thickness, supplied by JOVIR [8]. The diameter at the bottom part is 135 mm, and 60 mm at the top. Due to the fact that only the local behavior at the notch section was of interest, only the bottom part of the streetlight (at a height of 1.5 m) was tested with the high energy pendulum. A dead mass of 80 kg was fixed to the top part to reproduce the total mass and inertia of the complete streetlight. Three notch radii and three notch depths were used to weaken the streetlight base. The selected radii were: R=0.5, 1.5 and 4.0 mm. Taking into account that the thickness of the 3 streetlight is 3 mm, the selected depths were: D=1.0, 1.5 and 2.0 mm. The notch was machined along the perimeter, 200 mm above the bottom part of the streetlight. A view of the base of the streetlight, where the circumferential notch was made, is shown in Fig. 1. a) b) c) Figure 1. Detail of the notch, a) notch section detail view, b) detail of hollow streetlight’s wall, c) picture of a real notch machined at the streetlight base. Six streetlights were prepared with the combination of radii and depths shown in Table 1. The order shown in Table 1 corresponds to the order in which the streetlights were tested. Test specimens 2, 3 and 5 (in Table 1) have a constant depth (1.5 mm) and three different radii (0.5, 1.5 and 4.0 mm respectively). Test specimens 1, 3 and 4 (in Table 1) have a constant radius (1.5 mm) and three different depths (2.0, 1.5 and 1.0 mm respectively). These parameter combinations, keeping one of them constant, allow the influence on the impact behavior of the variable parameters to be seen more clearly. The reason for choosing these values is linked to the fact that the hollow streetlight’s wall thickness is 3 mm. 1.5 mm (one half of the wall thickness) was chosen as a reference value for depth and radius, and then, one value was chosen higher and one lower, both for the thickness and for the depth. Table 1. Radii and depths of the notches performed at the streetlight bases. Streetlight #6 does not have any notch. Superindex * indicates failure at the notch section. Radius (mm) 0.5 1.5 4.0 Depth (mm) 1.0 -- 4 -- 1.5 2 3 * 5 * 2.0 -- 1 * -- Note: Number 1-5 indicates the streetlight number. Due to the fact that the streetlights are manufactured by means of welding a curved steel plate along the longitudinal direction, the geometry of the transversal plane does not have a constant radius value. This means that direct machining along the external perimeter, using a 4 rotatory device, which would give rise to a truly circumferential geometry, cannot be performed. To solve this issue, the machining was completed by sectors, in an area where the external geometry had approximately a constant radius to guarantee a constant depth during the machining process. Depth measurements were taken, after the machining process, to guarantee that a reasonably constant notch depth was achieved. Other notch geometries, such as sharp V notches, have not been considered in this study. Sharp V notches together with the typical dynamic wind loads, would give rise to quick fatigue crack initiation. In fact, as will be seen later, from the three configurations with the same notch depth (1.5 mm) that with the larger radius (r=4 mm) has been chosen as the most appropriate to minimize the notch stress concentrations. 3. Impact tests A high energy pendulum was used for the impact tests. A steel block of 900 kg (“impactor” in what follows) was dropped from the adequate height to reach the resting line (see Fig. 2) at a nominal target speed of 30 km/h. The shape of the front part of the impactor is cylindrical. The distance from the resting line to the streetlight contact (“L” in Fig. 2) is 350 mm and the impact height (“Hs” in Fig. 2) is 400 mm, 200 mm above the notch height. The door to access the interior of the streetlight base was placed facing the impactor (originally, this access door was placed on the opposite side, but an unexpected failure of the door welding, at test #2, was observed, see comment on Section 4). Figure 2. Impact test scheme and definition of distances. For each test, two high speed cameras (1000 frames per second) were used, one recording a global view of the test and the other one with a detailed view of the notch area. Additionally, the horizontal deceleration vs time curve was recorded using an accelerometer (Endeveco, model 7264B-500T). The tests were carried out at CIDAUT Facilities [9] which has the EN 12767 impact test with accreditation issued by the Spanish accreditation body, ENAC [10]. The streetlight was fixed to a metallic plate (50 mm thick and fixed to the ground) by means of four H s L Impactor Resting line Streetlight Notch section 200 mm 200 mm Acces door to the streetlight 5 screws with hexagonal head M.16x35 grade 5.8 (UNE 135122 [11]) and 4 washers M.16 (EN ISO 7091 [12]). 4. Results and discussions Although it must be remembered that the acceleration profiles measured from a real crash test following EN12767 criteria are not comparable to the acceleration profiles obtained in the present test with a rigid steel block, this type of test (high energy pendulum) is typically performed in advance of the EN12767 test, as a preliminary, cheaper test. Figure 3 shows two pictures (a detail of the notch area and a global view of the test) of the high-speed recordings for each test. In the case of the streetlights failing at the notch section, the pictures correspond to those taken some milliseconds after the complete section failure. In the streetlights not failing at this section, the picture was taken after an approximately similar interval of time (around 12-15 ms after the impact). Streetlight - 1 Radius=1.5 mm Depth=2.0 mm Test number: E14-0855 Notch failure: YES Streetlight - 2 Radius=0.5 mm Depth=1.5 mm Test number: E14-0856 Notch failure: NO Streetlight - 3 Radius=1.5 mm Depth=1.5 mm Test number: E14-0857 Notch failure: YES Streetlight - 4 Radius=1.5 mm Depth=1.0 mm Test number: E14-0858 Notch failure: NO 6 Streetlight - 5 Radius=4.0 mm Depth=1.5 mm Test number: E14-0859 Notch failure: YES Streetlight - 6 NO NOTCH Test number: E14-0860 Notch failure: NO Figure 3. Pictures (detail and global view) of the impact tests. The first two tests were carried out with the access door on the opposite side of the impact. In test #1, failure was observed to occur at the notch section. In test #2, the local bending induced by the impact at the bottom part of the streetlight, with longitudinal tensile stresses at the door side, gave rise to an unexpected failure at the welding of the door with the streetlight column (which can be seen in the picture of the global view of test #2 and with more detail in Figure 4). The rest of the tests were carried out with the access door on the impact side of the streetlight. No more failures occurred at the welding. Figure 4. Detail of the unexpected failure at the weld seam of the access door (test #2). The horizontal decelerations were measured by an accelerometer inside the impactor. The six impact test results (acceleration vs time) are shown in Fig. 5 for the interval 0ms<t<30ms (using the legend: R=radius and D=depth). 7 Figure 5. Acceleration during impact tests. In Fig. 5, streetlights failing at the notch section are plotted with dashed lines, solid lines being used for the streetlights not failing at this section. Two key parameters have been extracted from the acceleration measurements, the maximum acceleration value, as an indicator of potential injuries to the vehicle occupants and the area under the acceleration curve as an indication of the energy absorption capability. It is clear from plots in Fig. 5 that for streetlights failing at the notch section, the complete failure occurs in the first 15 milliseconds, no decelerations being recorded afterwards. Streetlights not failing at the notch section showed a much higher energy absorption capacity with deceleration values different than zero beyond 30 ms after the impact. In Table 2, the maximum acceleration and area under the curve are summarized for the six tests. Streetlights 1, 3 and 5 (which failed at the notch section) have an average value of the maximum acceleration of 7.92 g (with a Standard Deviation of 0.89 g and a Variation Coefficient of 11.3%). Streetlight 2 had, as mentioned previously, an unexpected failure, and therefore it will not be to calculate the average values of parameters in Table 2. It can be clearly seen in Fig. 5 that the curve corresponding to streetlight 2 is different to streetlights 4 and 6 (those not failing at the notch section) with lower maximum acceleration values. The average value for the maximum acceleration from streetlights 4 and 6 (those not failing at the notch section) is 9.27 g (with a Standard Deviation of 0.29 g and a Variation Coefficient of 3.2%), 15% higher than in the weakened specimen tests. -2 0 2 4 6 8 10 12 0.000 0.005 0.010 0.015 0.020 0.025 0.030 acceleration (g) time (s) #1 (Notch R1.5-D2.0) #3 (Notch R1.5-D1.5) #5 (Notch R4.0-D1.5) #2 (Notch R0.5-D1.5)* #4 (Notch R1.5-D1.0) #6 (Without Notch) 8 Table 2. Maximum acceleration and area under the impact curves (0<t<0.030s). Maximum acceleration (g) Area (0 < t < 0.030s) (g·s) Notch Section failure Streetlight - 1 6.89 0.05498 YES Streetlight - 2 7.22 0.14203 NO Streetlight - 3 8.42 0.08447 YES Streetlight - 4 9.48 0.18544 NO Streetlight - 5 8.45 0.08039 YES Streetlight - 6 9.06 0.20354 NO The average value of the area under the curve for streetlights 1, 3 and 5 is 0.07328 g·s (with a Standard Deviation of 0.01598 g·s and a Variation Coefficient of 21.8%), while for streetlights 4 and 6 it is 0.19449 g·s (with a Standard Deviation of 0.01280 g·s and a Variation Coefficient of 6.6%), 62% higher. Streetlights with a notch depth of 1.5 mm or higher failed at the notch section, regardless of the notch radius value (with the exception of streetlight 2, which failed by the welding seam). Thus, the weakening of the streetlight could be defined by the following depth (1.5mm) and the highest notch radius value as (r=4.0 mm) in order to minimize the stress concentrations for the static and fatigue load cases. In Annex A, Finite Element Models are presented to calculate the stress concentration factors associated to the different notch geometries. In Annex B, the static strength and fatigue life are calculated for the final selected notch configuration (depth=1.5 mm and radius=4 mm). Results in Annex B show that the reduced section fulfills the design criteria (static and fatigue life). 5. Final remarks and conclusions Streetlights are support structures which are not specifically designed to absorb energy in an efficient way during a vehicle impact. From the results obtained, which have to be taken only as a first assessment of the potential behavior of streetlights under impact, the machining of a circumferential notch at the streetlight base reduces both the acceleration peak (by about 15%) and the impact energy which can be absorbed (about 62%). The excessive structural stiffness and structural strength of those streetlights already installed, allows the machining of the notch without reducing the structural stiffness of the streetlight and without reducing the structural strength below the required limits and as a result giving rise to reductions in the acceleration peaks. Although crash tests according to EN12767 standards are mandatory for the certification of the support structure, the preliminary results achieved using a high energy pendulum have shown that this notch would effectively reduce the acceleration peak values and the energy absorption. 15 In order to illustrate the calculated fatigue failure, based on the hypothesis of a continuously wind speed change (from 0 to vref) during the whole life, 25 years, of the streetlight, the number of changes per day associated, for example, to the vref =70 km/h would be 32 changes per day (daily, during 25 years), which looks pretty unrealistic to occur. In the case of vref =80 km/h (a much lesser probable average value for the reference wind speed), 1.5 changes per day, would be necessary, being also a quite unrealistic scenario. Thus, the selected notch geometry (depth =1.5 mm and radius=4 mm), does not seem reduce the fatigue life below a realistic threshold value. Of course, detailed fatigue life calculations should be performed for any other streetlight configuration (height).