cooling prize paper The uplift capacity of transmission pylon foundations

Transcription

1 The uplift capacity of transmission pylon foundations Frederick Levy, University of Southampton and National Grid UK Abstract The National Grid high voltage network requires urgent refurbishment to accommodate the demands of the 21st century. However, there is significant uncertainty surrounding the inservice uplift performance of transmission pylon shallow foundations under dynamic (wind) loading. This uncertainty has been driven by the apparent understrength of foundations when tested using current industry testing practice. A suite of fullscale tests on transmission pylon foundations was commissioned at a London clay site in Kent. The tests aimed to mimic inservice dynamic loading to reveal the mechanisms associated with rapid foundation uplift. It was shown that foundations founded on London clay mobilised significantly greater uplift capacities at smaller displacements compared with current industry practice. The results also suggested that aspects of current uplift design, namely failure mechanism geometry and ultimate limit state criterion, may also require revision. 1. Introduction In the UK there are approximately 22, high voltage (275/4kV) transmission pylons supported by a shallow foundation under each pylon leg. The loading of a transmission pylon will result in the foundation system being loaded in tension (uplift) or compression (bearing) (Figure 1). To prevent excessive buckling stresses, differential movement of the foundation system needs to be limited. During loading the compression foundation s displacement will be limited due to soil beneath it, whereas as uplifted foundation is pulled out of the ground it accumulates resistance before eventually shedding all load carrying capacity. The performance of the system therefore relies on a sound understanding of foundation uplift performance. Table 1: Foundation design parameters Soil strength Type Measurement Frustrum angle ( o ) Cohesionless Cohesive Strong Weak Strong Weak SPT N The high voltage network with incremental upgrades and refurbishments has been able to weather nearly half a century of changing demand and generation patterns. The recent 22 renewable targets (BERR, 28) in combination with saturation of capacity has necessitated a major program of restringing and uprating old cables (Clark et al., 26). This program will result in significant increases of loading on transmission pylon foundations systems due to larger cable sizes. The need for a better understanding of transmission Undrained shear strength (su) (N<2) 25 (su<49) 25 Source: National Grid, 24 Table 2: The design test from BS EN61773:1996 Test type Load steps of design Increment time (minutes) capacity (%) Design test 25, 5, 7, 8, 9 and at 1% Source: BSI, 1996 pylon foundation system behaviour is therefore required to facilitate this upgrade in a costeffective and timely manner. Recent studies undertaken by University of Southampton and field tests by National Grid have demonstrated that the design basis for transmission pylon foundations may not be reliable. It has been statistically shown that when subject to testing, half of the foundations tested did not reach their uplift design capacity (Figure 2). Yet the present failure rate of pylon foundations is extremely low, suggesting that there are additional Vertical loads Horizontal loads (tranverse and longitudinal components) Percentage of sample (%) Uplift resistance/design capacity (%) Figure 1: Transmission pylon foundation system and loading Figure 2: Sample of field tests (n=45) 24 ground engineering october 213

2 Load Backfill H=3m London Clay B=1.45m 2.5m Figure 3: L4M foundation Figure 4: Construction of an L4M foundation factors contributing to uplift resistance. Centrifuge model tests on transmission pylon foundations has shown that uplift resistance under rapid loading may be enhanced due to the development of negative pore water pressures (suctions) across the foundation base. It was shown that this increase in uplift resistance had a log linear relationship with the uplift velocity (Lehane et al., 28). These model tests were conducted using kaolin clay the preferred soil for centrifuge modelling work due to its uniformity and drainage characteristics. However, it exhibits a soft response in comparison to clays encountered in the field. Similarly, the close control in model tests may not be replicated in the field. These limitations prevent confident extrapolations from Table 3: Load schedule centrifuge modelling from being made. This paper aims to investigate pylon foundation capacity by conducting a set of fullscale rapid foundation uplift testing to compare the dynamic response to standard static testing practise. The purpose of the field tests was to introduce the effects of construction practices, tolerances and defects typical of those encountered in the field that are likely to influence the performance of a foundation subject to uplift. 2. Design and construction practice Uplift resistance (specified by National Grid in TS 3.4.) is derived from weight of the foundation and soil, contained within an inverted frustum (National Grid, 24) that extends to the surface from the base of the foundation (Figure 3). The frustum angle, and therefore the geometry of the failure mechanism, is governed by the insitu soil properties (Table 1). Foundation Test Backfill Base Uplift velocity (mm/s) Displacement w (mm) 1 1A Loose London clay 2 2A 2B 3 3A 3B 4 4A 4B Dense London clay London clay 35 London clay 5 Type 2 Type 2 1 Type 2 London clay Design test 5 5A Type 2 London clay 35 Figure 5: Load application The ultimate limit state (ULS) displacement criterion of shallow foundations is considered to be approximately 1% of foundation width (B) (BSI, 24). However, the ULS displacement (w) criterion of specified in TS 3.4. for all foundations is w = 1mm. This value is based on the assumption that it represents the elastic limit of the backfill in uplift (CIGRE, 1998). A truncated mass concrete pyramid base with a reinforced concrete inclined chimney is the most common foundation type for transmission pylons in the UK. The foundation is cast inside a large excavation typically backfilled using excavated or imported material compacted in 3mm layers. Foundation testing practice in the UK is codified by EN 61773:1996 (BSI, 1996). A summary of ground engineering october

3 Corrected cone resistance, q t (MPa) Sleeve friction (kpa) m = Type 2 31m+ = London clay 1m = London clay 3m = Type 2 31m+ = London clay 1m = London clay 1 1 Figure 6: Cone resistance profiles (Foundation 5) Figure 7: Sleeve friction profiles (Foundation 5) the incremental design test is presented in Table 2. This type of test has been routinely used to assess foundation capacity since the 196s (Parr and Vanner, 1962) and is the source of the uncertainty surrounding the in service uplift performance of transmission pylon foundations. 3. Field tests 3.1 Site layout and load schedule Five foundations were constructed at the Building Research Establishment London clay test site at Lodge Hill Camp in Chattenden during August 29. The foundations were designed and constructed to TS 3.4. (National Grid, 24) with uplift capacities of 42kN. To determine whether suctions occurred at field scale the foundations were installed with different base contact conditions, backfill material and uplifted at different velocities to w/b = 1% (mm). Two foundations were backfilled with compacted or loose London clay representing early construction practices (Parr and Vanner, 1962). The remaining foundations used compacted Type 2 a coarse granular material from recycled aggregate (Department of Transport, 29) and represented current practice. The load schedule in Table 3, in combination with the different forms of construction, allowed the isolation of each resistance mechanism. The performance of footings subjected to dynamic loading was scrutinised by using varying uplift velocities and displacements. A design test to BS EN61773:1996 (BSI, 1996) was also carried out to provide design test results for the L4M foundation type. Load was applied to each foundation stubs using an inclined hydraulic jack (Figure 4). The loaddisplacement behaviour of the foundations was measured using a load cell mounted above the hydraulic jack and linear voltage displacement transducers (LVDTs) on a reference beam. 3.2 Ground conditions Chattenden has been used extensively for foundation testing due to its deep (~3m) and uniform London clay strata. Testing was conducted over two weeks in July 212. Due to wet weather in the weeks prior to testing the top layer of weathered clay became soft requiring remediation with 4t of stone. Five 1m deep cone penetration tests (CPTs) were used to characterise the site and backfills. The measurements of pore water pressure from the CPTs through the backfills confirmed that water had significantly infiltrated the voids in the London clay and Type 2 backfills (Figure 8). The pooling of 1 1 water at the base of the excavations in combination with the rapid uplift velocities was expected to result in the development of an undrained condition (suctions) on the base of the foundations. The inundation of water in the compacted Type 2 granular backfills appears to have caused a significant loss in strength resulting in very low penetration resistances (Figures 6 and 7). The low strength of the granular backfill contrasts with the relatively stiff and homogeneous in situ London clay. The CPT results from Foundations 1 suggested Pore water pressure (kpa) Foundation 1 Foundation 2 Figure 8: Pore water pressure through backfills and in situ clay the same loss in strength was not experienced in the London clay backfills. The inferred values of undrained shear strength (su) and density (γ) of the London clay corroborated with previous observations (Butcher et al., 28). A summary of the in situ values from the CPTs at Footing 5 is presented in Table 4. These values were used for a set of effective and total numerical backanalysis studies. Table 5 presents a summary of the record uplift resistances (Q). Only the rapidly loaded foundations on Table 4: CPT results E backfill (kpa) E insitu (kpa) γ backfill (kn/m 3 ) 1.5 T2: 7,84 LC: 11,6 3 T2: 1,77 LC: 13,4 7,185 T2: 17.4 LC: ,3 T2:.3 LC: 18.5 γ insitu (kn/m 3 ) 18. T2: 31 LC: T2: 31 LC: 32 φ backfill ( o ) φ insitu ( o ) S u backfill (kpa) S u insitu (kpa) 32 T2: n/a LC: 5 31 T2: n/a LC: n/a 31,8 n/a 19.3 n/a 32 n/a ground engineering october 213

4 Normalised uplift resistance, Q/A (kn/m 2 ) Test 1A Test 2B v f =1mm/s, Lehane et al (28) v f =3mm/s, Lehane et al (28) Suction factor Field test Lehane et al (28) Normalised displacement, w/b (%) Normalised uplift velocity, v f B/c v Figure 9: Rapid load test results Figure 1: Suction factor Uplift resistance (kn), applied load (KN) Test 3B Test 4A (design test) Uplift resistance (kn) E u /s u =49 E u =184 MPa, Gasparre et al (27) E u =122 MPa, Gasparre et al (27) Displacement, w (mm) Displacement, w (mm) Figure 11: Design test results compared to pullout tests Figure 12 Total stress numerical results London clay mobilised resistances that were within 1% of the design capacity at ULS. The design test (Foundation 4 Test 4A) only reached ~5% of its design capacity at w = 1mm, consistent with previous observations (Parr and Vanner, 1962, Clark et al., 26). 4.1 Rapidly loaded foundations The rapidly loaded foundations on London clay exhibited an Table 5: Summary of results Foundation Test Q1mm (ULS) (kn) extremely stiff response, 9% peak resistance before w/b = 1%. The different uplift rates did not affect the loaddisplacement behaviour. The measured differences in uplift capacities between rapid tests was <1% at w = 1mm with a maximal difference of 5kN at peak. From Lehane et al. (28) it is inferred that the rate independence meant that full suctions occurred on the base and therefore peak capacity Q25mm (kn) Qpeak (kn) 1 1A A 2B 3 3A 3B 4 4A 4B A was limited by operational value of su of the London clay underneath the foundations. Figure 9 compares the field tests with results from the aforementioned centrifuge tests (Lehane et al., 28) with uplift capacities normalised to the foundation base areas. The peak capacities of the centrifuge tests are within ~2% of the field tests, which may be due to differences in the values of su between tests. Design capacity (kn) When compared using a suction factor, where the capacity difference between foundations with and without suction is normalised with the foundation area and su on the base (Qsucton Qno suction/sub2), the results from the Chattenden field test closely match that from the centrifuge model tests (Figure 1). It is proposed that the spline fit from Lehane et al. (28) is replicated for different values of consolidation coefficient (cv = khe /γw). 4.2 Static design tests results A design test (BSI, 1996) was carried out on Foundation 4 (Figure 11). Under the applied load increments the foundation performed poorly, reaching only 5% of design capacity at ULS. The difference between the rapid loading methodology and the design test was marked. It has previously been shown that peak loads are applied rapidly (Clark et al., 26). It may therefore be the case that the rapid test is more applicable and the design test is unduly conservative. ground engineering october

5 Figure 11 also shows the results from Test 3B; rapid uplift excluding suction. It is evident that the performance of Type 2 granular fill is extremely poor. Large displacements were required for the backfill to compact sufficiently to reach design capacity. Such a poor performance will require a reevaluation of the use coarse granular material, specifically Type 2, when used in excavations bounded by London clay. 5. Numerical modelling A set of numerical backanalysis studies were conducted to reproduce the loaddisplacement behaviour of the tested foundations. An axisymmetric foundation finite element model was analysed using OASYS SAFE (Oasys, 29) with total and effective stress soil models. The total stress soil model was used to examine the mobilised undrained shear strength and stiffness on the base of the foundation. The stiffness of the Chattenden field tests varied between Eu = 184MPa (w 2.5mm) and Eu/su = 49 (w = 5mm) (Figure 12). The former stiffness value corresponds to undrained triaxial tests on London clay samples from Heathrow Terminal 5 (Gasparre et al., 27). The large values of stiffness contrast starkly to the soft kaolin clay (Eu/su = 5) used in previous centrifuge model tests (Rattley et al., 28). The numerical back analysis also suggested that the operational value of su on the foundation bases was on average 41kPa, close to the CPTs values recorded on site suggesting that there may be merit in using such instrumentation to predict uplift resistances. The effective stress model aimed to capture the development of the granular backfill failure mechanism. The displacement vector plots of Figure 13 show that the uplift resistance at w = 1mm is solely due to the backfill overburden immediately above the pyramid. The failure mechanism at w = 5mm reaches the excavation interface but breakout was not observed on site. Both of these observations are contrary to the design methodology of the frustum method (National Grid, 24). 6. Impact The finds of the field tests have highlighted significant limitations of current design and testing practice. Current design practice makes use of the frustum mechanism, which due the geometry of excavations is assumed to breakout of the backfill into the insitu soil. However, it has been demonstrated through Figure 13: Granular backfill failure modes at w = 1mm and w = 5mm numerical back analysis that the failure mechanism may not occur when granular backfill is used to fill an excavation in cohesive soil. There therefore must be a careful reexamination of the failure mechanism used in design, with attention paid to combinations of backfill and insitu material. The field tests showed that the expectation that full uplift capacity will be mobilised within the specified ULS criterion of w = 1mm may be unreasonable. The field test results indicate that only those foundations that mobilised suctions had uplift capacities within 5% of design capacity at w = 1mm. To develop a more legitimate criterion a rigorous examination of foundation system performance (differential movement) with relation to the development of buckling stresses in the transmission pylon structure needs to be undertaken. Finally, the design test according to BS EN61773:1996 (BSI, 1996) appears to yield extremely conservative results in comparison to the rapid uplift methodology. Both methods seek to apply the maximum design load arising from transmission pylon design. Crucially, the rapid methodology does so within 2s, which may better replicate the dynamic loading of transmission pylon compared to the static incremental loading of the design test. Current statnamic testing of piles bears the closest relation to the dynamic testing used at Chattenden. 7. Conclusions The series of field tests on a number of fullscale L4M footings has confirmed that base suction may contribute significantly to footing performance. The results have also shown that the design uplift performance may be not reached (in general) before the specified ultimate limit state. This includes the performance of footings where suction developed. In the cases where suctions did not develop, the uplift performance of the footings was extremely poor. Also demonstrated was that current testing practice may have led to an undue underestimation of foundation uplift capacity due to the manner of test load application. The differences in observed uplift resistances due to rapid loading requires further consideration of the in service applied loading typically transmitted to foundation systems. This can be achieved through careful instrumentation and monitoring through foundation systems from their base upwards using accelerometers together with instrumentation to determine external loads and structural loads. With recent developments in data logging and transmission it would be relatively straightforward to fully instrument a high voltage transmission pylon from the base upwards. Through such endeavours a better design and testing rationale may be developed, leading to more sustainable refurbishment solutions. Acknowledgments The work presented forms part of a wider examination of shallow foundation performance, funded and supported by the National Grid. The author is grateful for technical advice and support from David Clutterbuck and Boud Boumeced of National Grid and University of Southampton professor David Richards. The foundations were constructed by Grid Line Foundations and testing was undertaken by ESG. The support of Andrew Hewitt of Lankelma is also gratefully acknowledged. References BERR (28) UK Renewable Energy Strategy. Department for Business Enterprise Regulatory Reform, London, UK. BSI (1996) EN 61773:1996. Overhead lines Testing of foundations for structures. BSI, London, UK. BSI (24) EN 19971:24. Eurocode 7: Geotechnical design, Part 1: General rules. BSI, London, UK. Butcher, A P; Powell, J M M; Kightley, M; and Troughton, V (28) Comparison of behaviour of CFA piles in London clay as determined by static, dynamic and rapid testing methods. In 5th International Symposium on Deep Foundations on Bored and Auger Piles (BAP V) (Van Impe, P & Van Impe, W F (eds)). CRC Press, Ghent, Belgium. pp CIGRE, (1998) Probabilistic Design of Transmission Line Structure Foundations. International Council on Large Electronic Systems, Paris, France, SC2 WG7. Clark, M; Richards, D J; and Clutterbuck, D (26) Measured dynamic performance of electricity transmission towers following controlled brokenwire events. International Council on Large Electronic Systems, Paris, France, Paper B2313. Department of Transport (29) MCHW Volume 1 Series 8, Road pavements Unbound Cement and Other Hydraulically Bound Mixtures. DoT, London, UK. Gasparre, A; Nishimura, S; Minh, N A; Coop, M R; and Jardine, R J (27) The stiffness of natural London Clay. Géotechnique, 57(1): 318. Lehane, B M; Gaudin, C; Richards, D J; and Rattley, M J (28) Rate effects on the vertical uplift capacity of footings founded in clay. Géotechnique, 58(1): 131. National Grid (24) TS Overhead line support foundations. NG, Warwick, UK. Oasys 29. Safe: Version 19. London, UK. Parr, R G; and Vanner, M J (1962) Strength Tests on Overhead Line Tower Foundations. Electrical Research Association, Leatherhead, UK, Report O/T28. Rattley, M J; Richards, D J; and Lehane, B M (28) Uplift performance of transmission tower foundations embedded in clay. Journal of Geotechnical and Geoenvironmental Engineering, 134(4): ground engineering october 213