10 PLAXIS Bulletin (S) - PLAXIS PLAXIS PLAXIS Editorial Nº...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: PLAXIS PLAXIS PLAXIS Editorial Nº 10 - MARCH 2001 Since the release of the Dynamics module last year users have been running dynamic In April the newly developed 3D Tunnel analyses. As this is a new module of Plaxis, program will officially be released . Many many users have to gain experience with it. months were spent in testing the program, Therefore, the Plaxis Users Association (NL) not only at the Delftech park office, but organised a well attended ‘Dynamics day’ with also by a group of Beta testers. The 3D guest speakers Prof. dr. ir. A. Verruijt and T.K. Tunnel program is specifically intended to Muller from IFCO. After these interesting model tunnels such as shield tunnels lectures, hands-on exercises were conducted (second Heinenoord Tunnel, see New in the afternoon to exchange practical Developments) and NATM tunnels. But the experience. Last January, during the course 3D Tunnel program allows other 3D ‘Computational Geotechnics’ at Berkeley situations to be modelled as well, such as University (USA) a one day Dynamics course the 3D excavation and installation of a was introduced. Some 30 participants from this diaphragm wall (M. de Kant, see Plaxis earthquake sensitive area were present. Practice). A short description of the 3D Tunnel program is provided in this bulletin. The analysis of flexible soil retaining walls is taken a step further. From Blum’s analysis and L Winkler spring type models to an analysis in 3D Tunnel program was presented at a users Plaxis. The approach of using a Finite Element method with an adequate soil model to analyse background of the 3D Tunnel program, a flexible soil retaining walls shows that some hands-on exercise was given on a simple interesting results can be obtained (Column example problem. Further more it was shown Plaxis bulletin Plaxis B.V. P.O. Box 572 2600 AN Delft The Netherlands E-mail: [email protected] ast year, in September, a Beta release of the meeting. Besides a lecture on the (theoretical) Bulletin of the PLAXIS Users Association (NL) Vermeer). that the PLAXIS 3D Tunnel program is certainly capable of other analyses beyond tunnelling. A case about the indentation of a tractor wheel about the role of OCR in the model. A practical excavation of a slurry wall were presented. Editorial implemented, questions have been asked in soft soil (University of Wageningen) and the IN THIS ISSUE: Since the Soft Soil Creep model has been example is given and preliminary conclusions Column Vermeer are drawn in ‘The role of OCR in the SSC model’ see Plaxis Practice. New developments Improved user services support Editorial staff: Recent activities Marco Hutteman, Plaxis Users Association (NL) Martin de Kant, Plaxis Users Association (NL) Peter Brand, Plaxis bv Plaxis Practice Jan Gabe van der Weide, Plaxis bv The role of OCR 12 Users forum 14 Agenda 16 Scientific Committee: Prof. Pieter Vermeer, Stuttgart University Partial geometry of NATM tunnel 1 Dr. Ronald Brinkgreve, Plaxis bv PLAXIS PLAXIS Column Vermeer FE-computations. We will consider a singleanchored wall for three different cases. ON SINGLE ANCHORED RETAINING WALLS We considered the geometry of Fig. 1, i.e. an The analysis of flexible soil retaining walls excavation depth of 10 m, an embedment of became possible through the work of Blum 2.5 m and an anchor at a depth of 2.5 m. The in the 1930s. Considering single-anchored anchor force was given a fixed value of or single-propped sheet-pile walls, he 100 kN/m. distinguished between two types of For all three different cases (A, B and C) the embedments: following soil properties were adopted: q free earth support q fixed earth support Submerged soil weight was used, as we ree earth support implies a relatively short F consider a water table at the soil surface, being wall with minimum embedment. Fixed not lowered at all, i.e. neither in front nor earth support implies a somewhat larger behind the wall. The excavation was done in embedment. According to Blum’s definition, three stages of construction: full fixity is achieved when the fixity moment 1 Installation of wall and excavation to a depth equals the field moment. of 2.5 m 2 Application of anchor load of 100 kN/m Blum’s design procedures for retaining walls 3 Excavation down to final depth with free or fixed earth support can be found in most textbooks. In the author’s opinion they Hardening soil model: Soil behaviour was constitute outstanding contributions to Soil simulated using the HS-Model of the Plaxis Mechanics. However, as Blum’s analysis involves code. For virgin oedometer loading, this neither the wall stiffness nor the soil stiffness implies an increasing tangent stiffness modulus it is bound to be inaccurate. As a consequence, according to one is now mostly using Winkler spring type with models. Unfortunately it is difficult to select appropriate spring constants and I would rather use the FE method. To assess the impact of where stiffnesses we decided to perform a series of adopted the exponent m = 0.5. Within the HS- is the major principal stress. We Model unloading-reloading is described on the Figure 1 Single-anchored wall with free earth support with 3 stages of construction: first excavation, anchoring and final excavation. In practice anchors will be installed just above the groundwater table. basis of Hooke’s law. Young’s unloadingreloading modulus for increments of stress and strain reads: Table 1 Stiffness parameters. where is the minor principal stress. For all analyses, the over-consolidation ratio was taken to be OCR = 1.0 and initial stresses were computed using Ko = 0.5. The HS-Model also allows for a specification of soil stiffness in drained standard triaxial tests. For all analyses, we used 2 PLAXIS PLAXIS The only difference between the stiff soil of The simulation of arching behind a flexible wall Cases A and B, and the soft soil of Case C relates makes the FEM superior to subgrade reaction to the stiffnesses. The stiff soil is simply a factor type models. In the latter case the spring will 15 stiffer than the soft soil, but the relation yield plastically as soon as eah is reached and is 1/1/4 for both soils. Moreover, active pressures will never reach smaller values both the stiff and the soft soil are conveniently given the same strength parameters. than eah= kah . ’. Fig. 2 clearly demonstrates z the significance of arching, as computed active Case A: Considering the FE-results for the earth pressures are well below the dashed line combination of a stiff soil and a flexible wall, for eah. It happens for flexible walls in stiff soils. one observes in Fig. 2a considerable wall bending up to about 5 cm. As a consequence, the active earth pressures reduce significantly; stiffness is the fixity of its base. There is a even below the classical minimum of e ah . Indeed, plots of stresses showed significant Figure 2 Single-anchored wall with free earth support. Another feature of a wall with low relative significant fixity moment! Here it should be arching between the anchor and the passive length using Blum’s design rules for a wall with pressure below the bottom of the excavation. no fixity at all. Due to the significant amount noted that we computed the embedment of arching and the base fixity, computed bending moments are small; approximately half the ones that would follow from Blum’s design rule. Case B: Typical Blum-type results are obtained when considering a stiff wall in a stiff soil (Case B). Below the anchor classical active earth pressures are reached. The passive ones are not fully mobilised, as we designed the wall for a factor of safety of 1.5 on the passive earth pressure. The base of the wall shows no fixity at all and bending moments agree well to the ones that follow from Blum’s analytical design procedure. Please note that the same earth pressures and bending moments would have been obtained for the combination of a soft soil and a flexible wall. In such a case we would have the same relative wall stiffness as for the stiff-stiff combination of Case B. Case C: I was amazed when considering computational results for a stiff wall in a soft soil. Despite the use of a factor of safety of 1.5, the passive earth pressure is nearly completely mobilised. It appears to be caused by an enlarged active pressure. Apparently, the soil is so deformable that wall displacements of about 5 cm are insufficient for a proper reduction of pressure on the active side. As a consequence of the high pressure a bending moment of nearly 300 kNm/m occurs. No doubt, this is well beyond the values that would 3 PLAXIS PLAXIS follow from Blum’s design analysis. 5 m is considered. Following Blum ’s design rules this would yield full base fixity, such that For a stiff wall in soft soil, I would also doubt the fixity moment equals the field moment. the results of subgrade reaction type Computational results for all three different calculations, as this method suffers from the relative wall stiffnesses are shown in Fig. 3. For difficulty of selecting proper spring constants. comparison, previous data for the shorter wall Realistic values would be required both for the are indicated by dashed lines. active and the passive zone; otherwise it is It appears from Fig. 3 that bending moments impossible to predict the high bending are only slightly reduced when increasing wall moments as obtained for a stiff wall in soft length. This is surprising as some textbooks soil. suggest a significant effect on the bending moments. Considering present computational Embedment: For studying the effect of Figure 3 Singleanchored wall for fixed earth support. Dashed lines indicate results for free earth support. data, we conclude that bending moments are embedment, we reconsider the wall of Fig. 1, in general not significantly reduced by but now the penetration depth is doubled. increasing wall penetrations. Present data Hence, instead of 2.5 m, an embedment of show, that the reduction of the field moment, as caused by the fixity moment, is more or less compensated by a slight increase of active pressure, as caused by the stiffening of the entire system. Deep penetration is neither of great import when considering displacements. Indeed, a significant reduction of displacements is only achieved for Case A. Conclusions: When considering a stiff wall in a stiff soil (Case B) typical Blum-type results are obtained. In this case classical active earth pressures will occur, at least below the anchor. Obviously, the passive ones will not be fully mobilised, if the wall is designed for a factor of safety equal of 1.5 on the passive earth pressure, as done in the present example. A flexible wall in a stiff soil (Case A) will result in considerable wall bending and low bending moments. The stiff soil transfers a large part of the active pressure by arching and the flexible wall gets a relative small pressure. A stiff wall in a soft soil (Case C) will result in high active pressures and, as a consequence, high bending moments. Finally we conclude that bending moments are in general not significantly reduced by increasing wall penetrations. P.A. Vermeer, Stuttgart University 4 PLAXIS PLAXIS New Developments monitored. Calculations of different construction phases are performed for the The Plaxis 3D Tunnel program is about to North bank. In a 3D finite element model (one be released. In the previous Bulletin it was symmetric half) the sub-soil, the Tunnel Boring explained why this first 3D Plaxis program Machine (TBM) and a part of the final lining were is devoted to tunnels. At the moment, modelled according to the ‘Grout pressure quite some engineering and research is modelling procedure’ (see Fig. 1). The sub-soil focused on tunnelling, both NATM and was schematised by means of 8 layers, with shield or bored tunnelling. Tunnelling their location and properties as listed in Table involves three-dimensional aspects that 1. All layers were modelled using the Mohr- cannot be analysed with conventional Coulomb model. The layers located under the methods. Hence, there is a demand for a tunnel were given a high unloading stiffness. 3D design model for tunnels. Nevertheless, creative users of the Plaxis 3D Tunnel The hydrostatic pore pressure distribution for program may find many other applications all layers was determined from a phreatic level in addition to the analysis of tunnels. at +1.0 m. n the past few months, beta-testers have I The 3D finite element model consists of 3440 used a pre-release of the 3D Tunnel program quadratic volume elements divided over a in practical applications. Some of these number of slices (see Fig. 2). Each slice is 3.0 preliminary results are presented in this m in the longitudinal tunnel direction. The TBM was modelled over 3 slices and composed of shell (plate) elements, with a flexural rigidity Table 1 Soil layers and parameters used in the Mohr-Coulomb model Layer Top Type m unsat kN/m3 E sat kN/m3 - c kN/m2 K0 kN/m2 EI = 50·103 kNm2/m, a normal stiffness EA = 10·106 kN/m and a weight w = 38,15 kN/m2. The radius of the TBM is 4.25 m and its MSL 17.2 27.0 0.0 0.58 0.35 m thick concrete tunnel lining was 6.5 0.47 modelled using volume elements with the 3.0 0.47 following properties: 5.0 0.45 24.6·106 2 1.00 Drained 16.5 17.2 0.34 3900 3.0 3 -1.50 Drained 20.5 20.5 0.30 29600 0.0 4 -5.75 Drained 19.0 19.0 0.31 18500 0.0 33.0 0.0 27.0 35.0 kN/m2, = 24 kN/m3, E = n = 0.2. 5 -10.00 Drained 19.5 6 -17.25 Drained 20.5 20.5 0.30 444000 0.0 36.5 6.5 0.50 7 -20.75 Undrained 20.0 20.0 0.32 119000 7.0 31.0 1.0 0.55 The tunnel boring process was modelled 8 -25.00 Drained 21.0 0.30 593000 0.0 37.5 7.5 0.56 according to the ‘Grout pressure modelling 21.0 19300 3.0 36.5 Undrained 16.5 0.30 3900 centre point is located at -12.3 m MSL. The 2.50 19.5 0.34 0.0 0.58 1 procedure’ as schematised in Fig. 1. Bulletin. In this article I will shortly present some results of a 3D calculation for the Second Heinenoord Tunnel, the first large-scale bored tunnel project under soft soil conditions in the South-West of The Netherlands. The situation at the Second Heinenoord Tunnel is described in various publications (see Figure 1 Modelling aspects in ‘Grout pressure modelling procedure’. References). The tunnel is formed by two tubes with outer diameters of 8.5 m, which were A front pressure was applied at the bore font bored under the river Oude Maas. In order to to support the soil. The front pressure is 140 gain experience with tunnel boring under soft kN/m2 at the top of the TBM and 259 kN/m2 soil conditions, the situation was extensively at the bottom. The TBM is conical. The tail 5 PLAXIS PLAXIS radius is 2 mm smaller than the front radius, m (TBM 3 slices, liquified grout zone 2 slices, which corresponds with a contraction of about tunnel lining 9 slices). The results of the 0.48% (0.16% per slice per phase). Behind the calculation at the end of Phase 14 are TBM grout is injected in the tail void. It is presented underneath. assumed that the grout remains liquified over 2 slices (6 m), which results in a grout pressure Fig. 2 shows the deformed mesh. This plot on the surrounding soil. The grout pressure is clearly shows the settlement trough at the kN/m2 at the ground surface, with a maximum settlement bottom. Behind the liquified grout zone the of about 22 mm. The results are quite realistic tunnel lining is activated and jack forces are and correspond reasonably well to the applied in backward direction (varying from measurements. This statement also applies to 125 Figure 2 Deformed mesh at the end of Phase 14 (deformations 50 times enlarged). 3365 at the top and 190 kN/m2 kN/m2 at the top and 6731kN/m2 at the the width of the settlement trough. bottom). Calculations using contraction only tend to In the initial situation, initial stresses are overestimate the width of the settlement trough, whereas calculations according to the ‘grout pressure modelling procedure’ give more realistic results. The deformations just above the tunnel lining are somewhat larger than at the settlement surface (max. 38 mm). Fig. 3 shows the shadings of total displacements. This plot confirms the above and clearly shows where the larger displacements occur (just above the lining). In this plot it can also be seen that the ‘buoyancy’ of the tunnel is relatively little, since the displacements below the tunnel are small. In addition to the displacements, the stresses can be visualised in the full 3D mesh as well as in individual and user-defined cross sections. From such plots the three dimensional arching around the tunnel can be viewed. There are also several possibilities to show the forces and deformations of the TBM and the tunnel lining, both in 3D and per cross section. The maximum bending moment in the TBM is around 100 kNm/m and the maximum bending moment in the lining is around 80 kNm/m. These values are also quite realistic. From the results it can be concluded that it is Figure 3 Shadings of total displacement at the end of Phase 14 (displacements 50 times enlarged). very well possible to calculate the three generated by means of the K0-procedure, dimensional effects around bore tunnels and using K0-values as listed in Table 1. The whole to accurately predict surface settlements using calculation is divided into 14 phases. In Phase the grout pressure modelling procedure. 1 the TBM enters the model in the first slice The above analysis took some 6 hours to and the process advances 1 slice in each phase. calculate on a Pentium III 500 Mhz with 768 MB In the final Phase 14 the TBM has advanced 42 RAM. 6 PLAXIS PLAXIS Improved User References: [1] Bakker K.J., van Schelt W., Plekkenpol J.W. (1996), Predictions and a monitoring Support Service scheme with respect to the boring of the Second Heinenoord In: As you may have read in the last Plaxis Geotechnical aspects of underground bulletin, the Plaxis company is growing. construction in soft ground, (eds: R.J. Mair Currently the staff consists of 10 people. and R.N. Taylor). Balkema, Rotterdam. pp. This means we can further increase the 459-464. quality and the range of products. In April [2] Jaarsveld E.P., Tunnel. Plekkenpol J.W., this will first show by the release of the 3D Messemaeckers van de Graaf C.A. (1999), Tunnel program. Obviously these new Ground deformations due to the boring products and the increasing group of Plaxis of the Second Heinenoord Tunnel. In: users (now over 2000 professionals !) require Geotechnical for good user services. Last year we employed transportation infrastructure (eds: F.B.J. a full time support engineer in order to Barends, J. Lindenberg, H.J. Luger, L. de keep up with the foreseen demand on Quelerij, A. Verruijt). Balkema, Rotterdam. profession support. infrastructure pp. 153-159. Tweede Heineneoordtunnel, verslag van W een groootschalig praktijkonderzoek naar Obviously such projects have a deadline. Hence, geboorde tunnels. Final report COB it is important that users can rely on a good committee K100. CUR / COB, Gouda. adequate HelpDesk. Not only on an operational [3] CUR / COB (1999), Monitoring bij de e realise that most support questions occur while working on a project. Validatie level, but also on a more scientific level. To ensure Groutdrukmodel, Meetveld Noord. Report urgent questions get the highest priority, we are of COB committee L520. CUR / COB, Gouda. introducing two different levels of support. [4] CUR / COB (1999), 3D [5] Bakker K.J. (2000), Soil Retaining Structures; development of models for structural Support service level 1: analysis. Dissertation (Delft University of This service level is free of charge and provides Technology). Balkema, Rotterdam. assistance by e-mail within reasonable response [6] Peters R., Safari B. (2000), 2D Modellering times, obviously depending on the support van het tunnelboor- en consolidatieproces. demand. No telephone and scientific assistance Internal is provided within this service level. report BSRAP-R-00004, Bouwdienst Rijkswaterstaat, Utrecht. [7] CUR / COB (2000), Toetsingsrichtlijn voor Contact HelpDesk: e-mail: [email protected] het ontwerp van boortunnels voor wegen railinfrastructuur. Final report COB Support service level 2: committee L500. CUR / COB, Gouda. The second level of support is to users who have a PLAXIS software support agreement. Ronald Brinkgreve, This agreement provides a professional support DELFT UNIVERSITY OF TECHNOLOGY service on operational as well as on a scientific & PLAXIS BV level, within 24 hours of normal working days. Contact HelpDesk: e-mail: [email protected] fax: +31 15 2600 451 telephone: +31 15 2600 450 For more information on the support agreement, please contact Plaxis BV. 7 PLAXIS PLAXIS Recent Activities courses. It is also good to realise that the ‘local’ courses are organised and lectured in Plaxis 3D Tunnel program cooperation with different local experts, local As this bulletin shows, the Plaxis 3D Tunnel university professors as well as professionals program is nearly finished. Currently the Plaxis from the engineering practice. This formula team puts the finishing touch to the 3D Tunnel creates tailor made courses, providing program and the final release is expected in theoretical and practical backgrounds on the April this year. The 3D Tunnel program is use of Plaxis. The courses in the Netherlands specifically intended to model shield, bored reflect the Plaxis philosophy with respect to and NATM type of tunnels. Hence the program lecturing and the usage of the Plaxis program. has special features such as a state-of-the-art The contents of the courses therefor are used tunnel designer, realistic grout pressure as the blue-prints of all international course. modelling, staged construction, Jointed rock model, etc. The userfriendlyness of the input, During the past course in Noordwijk, the the and Netherlands, we welcomed 41 participants robustness of the numerical procedures are in from virtually all over the world. Seventeen line with the Plaxis 2D software. This means different nationalities, including people from creating the finite element model and Japan, Jamaica, Mexico, Brazil, Israel, South prescribing the calculation phases are relatively Africa, just to mention a few. This is a clear simple, especially when compared to the indication of the true International character efforts it takes using general purpose FE codes. Plaxis is enjoying. automatic mesh generation Such codes are usually command driven and lack advanced options needed for the modelling of soil, structures and soil/structure interaction. With this new tool, Geotechnical 3D calculations can become generally applicable. Figure 2 Participants and lecturers in the Noordwijk course. So far some 10 new courses have been scheduled for this year. At the upcoming courses in Noordwijk, new lectures dedicated to both Dynamics and 3D modelling are presented (see also Agenda). Plaxis Practice Figure 1 The Plaxis 3D Tunnel Program RESEARCH PROGRAM ON THE IMPACT OF DIAPHRAGM WALL INSTALLATION. Courses VALIDATION OF A 3D-FE MODEL. In the past year 11 courses on Computational Geotechnics were lectured at different 1. Introduction locations all around the world. Six in Europe, The North-South metro line in Amsterdam will three in Asia, one in the US and one in the connect the northern and southern suburbs Middle East. From the agenda in this bulletin with the city centre (De Wit, 1998). For reasons you can see we plan to continue lecturing such of protecting the historic city centre and 8 PLAXIS PLAXIS restricting the disruption of city life, a bored tunnel will be applied that follows the street q measurements of stress in the trench (piëzometers on reinforcement cages). pattern as closely as possible and is lowered to a great depth. Consequently the underground The geotechnical profile is comparable with stations are at a great depth as well. The the locations of the future stations of the stations will be constructed in a building pit North-South line, see table 1. with 40m long braced diaphragm walls. One of the most important aspects in the design of Table 1. Geotechnical profile en the stations is the impact of construction on soil-parameters. historical buildings. Because knowledge on the Type of soil Top level NAP (m) impact of diaphragm wall construction on CPT (Mpa) surrounding buildings is only limited, it was Fill (sand) +2,0 10-15 decided to carry out a research project, Clay -1,0 0,5 consisting the following four phases: Peat -3,5 0,5 1 prediction of the impact with a 3D FE- Clay, silt -7,0 0,5-1,0 model; Peat -13,0 1,5 2 full scale test; Sand -14,0 1) 8-30 3 interpretation of test results; Silty sand -17,0 1-5 4 validation of the 3D FE-model. Clay (eemklei) -25,0 1-5 This paper focuses on phase four. Sand -28,0 10-30 Silt -42,0 3 2. Full scale test 1) = The test was carried out at the construction site Foundation layer of ancient buildings on wooden piles of the Mondriaan Tower, where diaphragm wall panels are applied as foundation elements as 3 FE - validation calculations well as building pit walls. Figure 1 gives an The main goal of the validation calculations is impression of the test site. During the different developing a model which can be used to stages of construction of 5 panels, a monitoring predict the soil displacements during the program was conducted consisting of: construction of a diaphragm wall. The Fig. 1 Impression of the test-site calibration was carried out, mainly by changing the soil parameters within a certain bandwidth. (extensometers, inclinometers); q measurements of vertical and horizontal displacements in the surrounding soil q This calibration procedure has resulted in a settlement and bearing capacity tests on “best-fit” piles; calculations were made with the official calculation (BF). Additionally, “North/South-line parameter-set” (NS), which has also been used in the 2D tunnel- and building-pit calculations. In both calculations, BF and NS, the hard-soil model is used. Element mesh and modelling procedure For the 3D-calculations a preliminary 3D-version of PLAXIS (6.4b) has been used. The mesh is build of 3D-wedge-elements with 15 nodes and 6 Gauss-points. For the calculation of one panel a mesh with approx 5,000 nodes has been used (fig. 2). Only a quarter of the panel is modelled, which means that there are two planes of symmetry: in plan view x- and zdirection. 9 PLAXIS PLAXIS Fig. 2 Finite element mesh The construction of a single diaphragm wall panel is modelled using the following stages: 1 Excavate the single trench by switching the soil elements off and, simultaneously, =d. applying the bentonite pressure ( b is hardened. Above the hardened concrete, the wet = d . ) acts on the concrete pressure ( c faces of the trench ( c c = volume weight of wet concrete). The lateral pressure in step 2 can therefore be described by a bilinear relation: were hcrit is a critical depth which depends on the concrete placing rate, cement type, temperature etc. (Lings et al 1994). In case of a constant rate of concrete placing, to is equal b. During installation the stresses in the soil will initially decrease (stage 1) and, subsequently, increase due to the wet concrete pressure (stage 2). The stresses decrease exponential with increasing distance from the trench. Among these stress-paths the stiffness of the soil varies strongly. The calculations where therefore carried out using non-linear material models: the hard soil model (HSM) and the soft-soil model (SSM). Only the elements representing the clay and peat layers are modelled as undrained. Calculation results for a single panel Figure 3 shows the horizontal effective stresses, at a depth of NAP-15m, during the different stages (1 and 2). The initial horizontal stress is about 50 kPa. During excavation the stress at the centre of the panel decreases to about 25 kPa. Due to horizontal load transfer (arching), at the edges of the panel, the stress increases and exceeds K0 situation. After concreting, the stress at the centre increases significantly. At the corner a decrease of effective stress occurs. The calculations have shown that load transfer not only acts horizontal but also in vertical Fig. 3 Horizontal effective stresses in different stages b) on the faces of the trench (d = depth, b = volume weight of bentonite). 2 Fill the trench with concrete by increasing direction to relative stiff layers. The displacements in layers beneath those stiff layers therefore decrease and are no longer of the lateral pressure. influence on displacements of the top layers. Directly after pouring the trench, from Therefore the influence of the panel-depth in bottom up, the lower side of the concrete a layered subsoil with stiff layers is very small 10 PLAXIS PLAXIS which was an important conclusion of the prediction calculations. The model was calibrated on the displacements in the concrete-phase (stage 2), because displacements during the excavation phase (stage 1) were to small. Figure 4 shows the horizontal displacements at the centre of the panel, at approx. 2m from the trench. The maximum displacement occurs in the Holocene top-layers, particularly the peat layer at NAP-6m. The BF-calculation is in good agreement with the test-results. Regarding the top layers, the NS-calculation is also accurate, Fig. 4 Horizontal displacements, middle of the panel but the horizontal displacements at lower layers are over predicted. The computed vertical maximum displacements of the sandlayer at a level of NAP-15m (foundation layer of wooden piles) are for both the BFcalculation and the NS-calculation in good agreement with the test-results (fig 5). At the test-site a settlement bowl occurs with a maximum settlement at a distance of 3m from the trench. This is in good agreement with the NZ-calculation, although the calculations shows a much wider settlement bowl. Influence of panel size and installation sequence At the test-site 5 panels were monitored, with different sizes and shapes. In contrast to the Fig. 5 Vertical displacements on level NAP -15m calculations and literature, the panel width was of little influence on the displacements. The reason for this discrepancy has probably been the installation sequence of the panels. The small panel (b=2,7m) was excavated in a undisturbed situation whereas the wide panel (b=6,4m) was excavated between two former installed panels. Those adjacent panels causes a reduction of displacements because of: q hardening (due to the installation of the adjacent panels the soil is overconsolidated); q load transfer to adjacent stiff concrete walls. To confirm this interpretation, the described situation was computed with the calibrated model (NZ-parameters). Figure 6 shows the Fig. 6 Influence of panel size and sequence of installation on horizontal displacements horizontal displacements of undisturbed panels with different widths, and a wide panel 11 PLAXIS PLAXIS between to installed small panels. Although hardening was not calculated due to the undrained behaviour, the results are in The role of OCR in the SSC Model agreement with the expectation. The Over-Consolidation Ratio (OCR) is Conclusions defined The diaphragm wall research project has preconsolidation stress enlarged the understanding of the impact of in-situ stress: OCR = D-wall installation on surrounding soil. Main parameter that indicates the amount of conclusions are: overconsolidation of the soil. As long as the as the ratio p between the p and the effective / ’yy. OCR is a state soil behaviour is stiff. This applies to the ground deformations were relatively unloading as well as to reloading. As soon as the existing preconsolidation stress is displacements in the soft top layers; passed by primary loading of the soil, the the width of a diaphragm wall panel is of soil great influence on ground displacements. preconsolidation stress further increases Due to vertical load transfer, the influence along with the effective stress level. In that of the panel-depth in a layered subsoil with case the soil is said to be in a state of stiff layers, is relatively small; normal consolidation, which would imply the computed displacements with the OCR=1.0. This is exactly what happens in calibrated 3D-FE model are in good Cam-Clay type of models, like the Soft Soil agreement with the test -results. q preconsolidation stress, i.e. OCR>1.0, the small, with exception of the horizontal q effective nearby wooden pile foundations; q the installation of a diaphragm wall in a layered subsoil has no significant impact on q model in Plaxis. stress behaviour is is soft, below whilst the the Ingenieursbureau Amsterdam A North/South Line Design Office, Plaxis. In that respect the meaning of the OCR- Amsterdam, The Netherlands value is similar, although the transition from similar behaviour can be observed when ing. M. de Kant using the Soft Soil Creep (SSC) model in the stiff reloading behaviour to soft primary Literature loading is more gradual (see Fig. 7.13 of the Lings, M.L. & C.W.W. Ng. & D.F.T. Nash, 1994, Plaxis Material Models Manual [1]). In order for The lateral pressure of wet concrete in the preconsolidation stress to ‘follow’ the diaphragm wall panels cast under bentonite, effective stress level, time is needed. Hence, Proc. Instn. Civ. Engrs. Geotech. Engng, 163- when loading the soil very quickly, the OCR- 172. value can (temporarily) become less than 1.0. De Wit, J.C.W.M., 1998, Design of underground On the other hand, if the load remains stations on the North/South line, Proceedings constant, the creep process continues in time, of the World Tunnel Congress, Sao Paolo. which results in a gradual increase of the De Wit, J.C.W.M., J.C.S. Roelands & M. de Kant, preconsolidation stress and OCR (without Full scale test on environmental impact of physical loading). The latter process can be diaphragm in conceived as ‘ageing’. The idea that ageing of would Amsterdam, wall trench Geotechnical excavation Aspects lead to an increase of the Underground Construction in Soft Ground, IS preconsolidation stress was introduced by 1999 Tokyo Japan, Balkema Publ. Bjerrum in his creep formulation [2]. This idea is also adopted in the SSC model. In the SSC model the increase of the creep strain in time 12 PLAXIS PLAXIS (= creep strain rate) depends (in addition to the west of The Netherlands) the initial OCR-value *) on the ratio of the is, per definition, equal to 1.0. On the other effective stress and the preconsolidation stress, hand, even very soft soils often show a i.e. on the inverse of OCR. Hence, starting from preconsolidation stress that is over 20 kPa an initial effective stress state, the initial creep beyond the in-situ stress level. Particularly in strain rate depends on the initial OCR-value. the top layer (just below the ground surface) modified creep index, crust forming may occur, which is associated The latter is not very known among Plaxis with relatively stiff and strong behaviour. The users. Since practical situations always involve reason for this can be drying of the soil, effective stresses from the very beginning, the variations of the phreatic level, temporary creep process (settlement) starts immediately loads, temperature changes, etc. Nevertheless, without additional loading, whereas the these soil are still considered to be ‘normally ‘settlement velocity’ depends on the OCR- consolidated’, but the actual OCR-value is often value. For reasonable combinations of SSC higher than 1.0. parameters, the use of a default initial OCRvalue of 1.0 in the K0-procedure may lead to excessive initial settlement velocities. Hence, layer under standard boundary conditions with the initial OCR-value needs to be selected with properties as listed in Table 1. The parameters care. An initial OCR-value larger than 1.0 is Figure 1: Time-settlement curves for different initial values of OCR Let us consider, for example, a 10 m thick clay are arbitrary, but realistic for a normally generally recommended. consolidated clay. Initial stresses are generated It is generally said that for ‘normally with the K0-procedure, using different initial consolidated soils’ (like the soft clays in the values of OCR (OCR0). On using different OCR0values, Plaxis proposes different K0-values, but all K0-values are reset to their original value of K0nc=0.703. For all values of OCR 0 a drained calculation is performed for a total time of 1000 days, just to let the soil creep under its self weight without additional loading. The results of these calculations are presented in Fig. 1. From Fig. 1 it can be seen that there is a remarkable difference in settlement, which is particularly caused by the initial inclination of the time-settlement curve, i.e. the initial creep rate. For OCR0=1.0 the initial creep rate is quite unrealistic. Considering rather soft ‘normally consolidated’ soils, it is quite realistic to have a settlement of 0.05 m a year, decreasing down to 0.01 m a year when the soil hasn’t been disturbed for a some years. In the above situation this is well reflected by the choice of Table 1. Properties of arbitrary clay layer, OCR0=1.4. At t=1000 days the inclination of all modelled with the Soft Soil Creep model wet 17 * * * 0.02 0.10 0.005 c 0.15 1.0 K0nc 26.0 0.0 curves is almost the same, except for 0.70 OCR0=2.0, which would suggest that the actual OCR’s have increased to almost the same value. kN/m3 Evaluation of the actual OCR-values at t=1000 13 PLAXIS PLAXIS days gives OCR 1.8 for all cases, except for Users Forum OCR0=2.0. In the latter case OCR has hardly increased. Question: I have some difficulties in generating the initial From the above results it could be concluded stresses in my project. My project is a tunnel that an initial OCR-value of 1.4, to be used in 200.0 m underground and is not including the the K0-procedure, would be a good choice. This ground surface; otherwise the model will be could indeed be said for the above example too big. Therefore, the initial stresses at the and perhaps for other cases, but it cannot be top of the model are not zero and should stated in general. It is a good habit for a include the overburden above the top of the practical application to simulate a certain creep model. Could you please give me a hint solving period in a drained calculation without this problem? additional loading and to evaluate the initial OCR-value. Please note that when using initial Answer: OCR-values larger than 1.0, Plaxis will propose If the model becomes too large you can K0>K0 nc, whereas for ‘normally consolidated eventually omit the upper part, but you have soils’ it is recommended to reset K0=K0nc. If the to compensate for the missing soil weight resulting settlement velocity is too high and otherwise non-realistic stresses will be the other parameters of the SSC model have generated. To include an overburden pressure been properly determined, then the initial value you have to do the following: Create a thin of OCR should be increased (or the modified layer with thickness h at the top of your model creep index * should be reevaluated). representing the omitted soil (see Figure 1). The weight of the soil in the thin layer is higher In conclusion, the initial OCR-value in the Soft than the omitted soil and equal to ( Soil Creep model needs to be selected with ( real * hreal / hvirtual. For example in Figure 1, care. In principle, the initial value depends on ( virtual the ratio between the initial preconsolidation this way you can generate realistic initial stress and the effective in-situ stress (which is stresses by means of the K0 procedure. virtual = = 18.0 * 130.0 / 1.0 = 2340.0 kN/m3. In generally slightly larger than 1.0), but it should be realised that the initial OCR-value also determines the initial settlement velocity. A first estimate could be OCR=1.4, but it is advisable to simulate the creep process and adapt OCR if necessary. RONALD BRINKGREVE, DELFT UNIVERSITY OF TECHNOLOGY & PLAXIS BV References [1] Brinkgreve R.B.J., Vermeer P.A. (1998), PLAXIS Finite Element Code for Soil and Rock Analysis, Version 7, Part: Material Models Manual. Balkema, Rotterdam Figure 1: Replacement of 130.0 m of soil by 1.0 m of soil with a higher density. [2] Bjerrum L. (1967), Engineering geology of Norwegian normally consolidated marine Question: clays as related to settlements of buildings. Why are the stresses of non-porous materials Géotechnique 17 (2), pp. 81-118. not shown and how can I visualise them anyway? 14 PLAXIS PLAXIS Answer: Non-porous material is general used for structural purposes like concrete, and stresses can therefore become quite large. As a result, stresses in the soil cannot be viewed in detail. By excluding the stresses in structural elements, the soil stresses remain clearly visible. The stresses in non-porous material are calculated though and can be viewed in tables and cross-sections. However, there is a trick to view the stresses in non-porous material: Calculate the problem and if all the calculations are finished and you want to have a look at the stresses in output then save the calculation data by pressing the save icon. Then go directly to Input and open the concrete material data set. Change <non-porous> to <drained> and exit the material data set. Save the Input data and go directly to Output by pressing the Output icon. In Output the stresses in the nonporous material set are now visible. Output is tricked in believing that the non-porous material set is drained and shows the stresses. 15 PLAXIS PLAXIS ACTIVITIES 03-06 JANUARY, 2001 27-31 AUGUST, 2001 Short course on Computational XVth ISSMGE International Conference Geotechnics (English) on Soil Mechanics and Geotechnical One day Dynamics course included Engineering (XV ICSMGE) Berkeley, California, U.S.A. Istanbul, Turkey 22-24 JANUARY, 2001 01-03 SEPTEMBER, 2001 Course on Computational Post-conference event (XV ICSMGE) Geotechnics (English) Short course on Computational Noordwijkerhout, The Netherlands Geotechnics (English) Istanbul, Turkey 19-21 MARCH, 2001 Course on Computational WINTER, 2001 Geotechnics (German) Short course on Computational ‘Finite Elementen Anwendungen in der Geotechnics (French) Grundbaupraxis’ ‘Pratique des éléments finis en Stuttgart, Germany Géotechnique’ 20-22 MARCH, 2001 Paris, France Short course on Computational Geotechnics (English) 20-23 JANUARY, 2002 Course on Computational Kuala Lumpur, Malaysia Geotechnics (English) 26-28 MARCH, 2001 Noordwijkerhout, The Netherlands International course for experienced Plaxis users (English) 24-27 MARCH, 2002 Noordwijkerhout, The Netherlands International course for experienced Plaxis users (English) Noordwijkerhout, The Netherlands 02-04 APRIL, 2001 Course on Computational Geotechnics (English) ‘Numerical methods in Geotechnical Engineering’ Eynsham Hall, Oxfordshire, United Kingdom For more information on these 23-25 APRIL, 2001 activities please contact: Short course on Computational Geotechnics (Arabic/ English) Plaxis bv Cairo, Egypt P.O. Box 572 2600 AN 06-09 AUGUST, 2001 DELFT Short course on Computational The Netherlands Geotechnics (English) Tel: +31 15 26 00 450 One day Dynamics course included Fax: +31 15 26 00 451 Boulder, U.S.A. E-mail: [email protected] 16 ...
View Full Document

This note was uploaded on 11/19/2012 for the course STRUCTURE 2121 taught by Professor Ed.jk. during the Spring '12 term at ENGECON University.

Ask a homework question - tutors are online