Experimental justification of the stability of the floating unit

One of the main tasks arising when installing a floating hydroelectric power unit on a foundation without preliminary excavation is a thorough justification of the shear stability and bearing capacity of the "floating hydroelectric power unit foundation" system on a complex geological massif. Failure to take into account these factors can lead to serious consequences during the landing of the structure in the target and further operation. It should be emphasized that this problem still includes a number of difficulties and does not always allow obtaining exact solutions in a volumetric setting. Based on the selection of a wide range of model materials, bases of various capacities were modelled for four models. In this case, the shear real characteristics of alluvial soils and their change after reinforcing cementation were taken into account. The studies were carried out on 4 models under static loads with bringing them to destruction. The models reproduced the real geological conditions at the base of the block, simulated deformation, and shear characteristics. Indicator diagrams of displacements, damage patterns, and generalized safety factors for bearing capacity were obtained. Model tests have shown that reinforcing cementation reduces not only the values of horizontal and vertical displacements of structures but also leads to a significant increase in the safety factor.


Introduction
The main problem that arises when installing a floating hydroelectric power plant (HPP) unit directly on the base is the need for careful consideration of the physical and mechanical characteristics of the soil that composes it [1][2][3][4]. Underestimation of the engineering and geological conditions of the installation area often causes the emergency state of the structure during the construction period and during the operational period.
All this makes it necessary, along with computer calculations according to special programs, to conduct model studies to substantiate the reliability of the floating block operation on a complex soil mass. In recent years, due to the difficulties in organizing research on physical models, this issue has been practically ignored, which often leads to undesirable consequences.
As the initial data for assigning the design characteristics of the foundation, the engineering and geological conditions in the sections of the Angara -Yenisei cascade were considered. Options for installing a floating block on a base made of alluvium, which is most common in this area, were studied.
Previous studies [5][6][7][8][9][10][11] have shown that base deformations caused by the displacement of the structure model under the influence of applied loads when bringing the model to failure can spread to a considerable depth. In this case, the physical and mechanical characteristics of rocks lying below those located directly under the structure can affect the nature of destruction and the magnitude of the breaking load. In this regard, it seems necessary to model the underlying alluvium rock as well. In the sections under consideration, the underlying layer is represented by a rock of intense weathering (zone "A"), the characteristics of which are given in Table 1.
Separations of the destroyed rock within zone "A" in the upper part are much less than in the lower one. In this regard, for more accurate reproduction of the geomechanical properties of the base, it was decided to model the underlying rock in two tiers: the upper of them is transitional from alluvium to destroyed rock with larger parts. Considering that the deformative characteristics of the base, along with the shear strength characteristics, can be of decisive importance when modelling the base. This significantly complicated the task of selecting model materials since it required linking different layers of the model base taking into account three factors: c, φ, and E (respectively, adhesion, angle of internal friction, and deformation modulus).

Initial data for modeling loads
Shear tests of the floating block models were carried out under static loads. The own weight of the unit, equipment, weighing, and anti-filtration back pressure was taken into account; hydrostatic loading in the upstream and downstream.
Loads were calculated based on the following assumptions: it is considered as the worst case of repair of the draft pipe from the point of view of ensuring the stability of the block in shear; adopted head Н = 15m at a relative elevation DSL 5.5m; relative lower elevation of the block bottom -7.25 m; the relative upper elevation of the block is taken on the basis of a margin above the HWL of 3m and is 16.25m; block dimensions in plan 28 x 28 m (for the layout of a hydroelectric power station building with horizontal turbines and a bell pipe); the height of the block at marks -7.25 m and 16.25 m is 23.5 m ( Figure 1); the given filtration diagram is shown in Figure 1; the given hydrostatic pressure diagram is shown in Figure 2; the volume of concrete is 7322m3, the weight of concrete at γ_c= 2,5 t/m3 is 18305 t; weight of water above the block elements for the case of repair of the exhaust pipe is 4298 t; equipment weight is 157 t; Prismatic strength of concrete is R_pr=20,0 MPa (for 180 days).   The physical and mechanical characteristics of the materials used for the model are given in Table 2.
The selected compositions for different horizons of the models' foundations allow, as can be seen from Table 2 to model rather well the ratio of the strength and shear characteristics of the "block-base" complex while maintaining the ⁄ ratios for concrete of nature and the model, as well as the ratio of the normative characteristics of the block and bases of nature and model.

Test bench and model loading systems
Models of the floating block with their bases were tested in a stand (Figure 4), which is a rigid welded structure with internal dimensions: length -320 cm, width -50 cm; height -185 cm. The dimensions of the stand allow placing two models in it at the same time. The load-bearing floor (2) is raised from the bottom of the standby 80 cm. The bases of the models (3) are laid on it, and the block models (4) are installed. Through the holes in the load-bearing floor, the rods of its own weight (6) are passed, laid down in the model when it is cast in the center of gravity of the block, and the lower ones -to the loading beam of its own weight (7), with the help of two hydraulic jacks (8), located symmetrically relative to the longitudinal axis block, the model is loaded with its own weight.
The model is loaded with its own weight using a press (9), the pressure in the gravity system is controlled by a pressure gauge (10). A shear force is applied to the block model using a horizontally mounted hydrostatic load jack (11) and a die (12). The model is loaded with a hydrostatic load using a press (15), the pressure in the hydrostatic load system is controlled by a pressure gauge (16).

Measuring equipment
When carrying out experiments on the models, displacements were measured using dial indicators. The indicators were installed at the points shown in Figure 5. As can be seen from Figure 5, according to the readings of the indicators, in addition to the absolute values of the displacement values, it is possible to estimate the angles of rotation of the model in different planes and control the symmetry of the model.

Scheme of the experiment
All models were tested with bringing to failure in order to determine the coefficient of shear stability: Were: Р . .load at which destruction or loss of stability of the model occurs; Р . . -(calculated) -design load on the model; Р . . is recorded on the hydrostatic pressure gauge at the moments of destruction, that is, the impossibility of further increasing the load. Prior to the experiment, the model was preloaded by 0.2 . at 0.2 . In order to compress it, eliminate possible backlash in loading systems.
The model was tested at a constant value of its own weight . . It was taken into account that we have the initial load in the form of the own weight of the block and the loading beam, equal to 41 kg, which is 0.11 . . Further, increase in the load was carried out in steps of 0.1 . ; the last stage is 0.09 . . The hydrostatic load increased up to destruction in steps of 0.1 . . At each stage of the load, both from its own weight and from hydrostatics, the displacements were measured: the first -immediately after loading, the second -after the time delay required to stabilize the deformations.  Upon reaching the calculated value of its own weight and subsequent loading of the models with hydrostatic load, the models on unreinforced alluvium show nonlinearity from the moment the hydrostatic load increases and occurs during the entire loading process.

Results and Discussion
For models on a reinforced base, sections of the diagrams can be distinguished that are close to linear at (0.4 ÷ 0. 5) . . With a further increase in hydrostatic pressure, the indicator diagrams also become nonlinear.
Upon reaching . , the magnitude of horizontal displacements on an unreinforced base is 3 ÷ 5 times higher than on a fortified one. For example, on models No. 1 and No. 2, they reached 81 mm and 66 mm, respectively, and on models No. 3 and No. 4, respectively, 17 mm and 25 mm (Figure 7).

Pictures of destruction
Analysis of the fracture patterns allows us to note that the loss of the bearing capacity of all models was due to a shear that took place with an increase in the hydrostatic load.
Model No. 1 (thickness of the alluvium layer 5m) collapsed at 1.1 . . a shift was noted along the unreinforced alluvial base, accompanied by sedimentation of the lower part of the block and soil uplift in the downstream. During disassembly of the model and removal of the block, the surface shear track had the character of a flat smoothed area with a slope towards the downstream. The depth of the track at the lower edge of the model was about 1 cm.
During the unloading of sand and expanded clay after the experiment, deformations of the model of the underlying rock of zone A2 in the form of a rectangular track 0.5 ÷ 1 mm deep were also noted.
Model No. 2 (thickness of the alluvium layer 15m) collapsed at 1.7 . The shear pattern is similar to model No. 1. A similar footprint is noted in the sandy base. However, the significant thickness of the alluvium layer compared with model No. 1 prevented the destruction of the underlying rocks A1 and A2. No deformations or destruction were recorded in these parts of the A1 and A2 rocks. Model No. 3 (reinforced alluvium 5m thick). Loss of bearing capacity was due to shear at 2.2 . During failure, a small uplift of soil in downstream was noted. The shift took place along the bottom of the fortified part of the alluvial base model, just above the contact plane "fortified soil -rock of zone A2". At the same time, the model of the rock base did not collapse (Figure 7).
Model No 4 (reinforced alluvium 15m thick). The model collapsed at 1.9 . The destruction occurred similarly to model No. 3, which was also found during disassembly and inspection of the model (Figure 7).