Investigation on Hydraulic Fracturing and Flexible Anti-Hydrofracturing Solution for Xiaowan Arch Dam

Abstract

Understanding hydraulic fracturing in concrete super-high arch dams is vital for the implementation of safety measures on the bearing surface. In this study, we conducted tests on hydraulic fracturing for the Xiaowan arch dam (294.5 m) to analyze concrete behavior at cracks under various stress conditions. The risk of hydraulic fracturing near the dam heel was identified without compressive stress. Addressing this, we propose a flexible anti-hydrofracturing system using GB sealing material and a spray polyurea coating. Simulation tests on three schemes: ‘3 mm GB plate + 4 mm polyurea’, ‘1 mm GB glue + 5 mm polyurea’, and ‘7 mm polyurea’ showed effective prevention of hydrofracturing at concrete crack openings of 5 mm, 8 mm, or 10 mm under 300 m water pressure. Field tests supported ‘3 mm GB plate + 4 mm polyurea’ and ‘7 mm polyurea’ as optimal solutions for dam sections. Implementation involves a protective block layout with ‘3 mm GB plate + 4 mm polyurea’ on blocks and ‘7 mm polyurea’ in interval zones and corners. Since 2008, maximum leakage, including rock foundation, has remained minimal at 2.78 L/s under regular water levels. These insights aid similar concrete dams in optimizing safety systems.

1. Introduction

Arch dams find widespread application globally, offering substantial economic benefits alongside inherent risks. The complexity of stress distribution, particularly evident in super-high arch dams with heights around 300 m, complicates crack assessment within the dam heel. Subsequent to impoundment, the emergence of surface cracks can lead to hydraulic fracturing failures due to high water pressure, entailing intricate and costly repairs [1,2,3]. Illustrating this, the Sayano-Shushenskaya arched gravity dam (H = 245 m) encountered numerous cracks during construction due to temperature control oversights. In its seventh operational year, the No. 35 dam section witnessed an abrupt escalation in crack propagation upon impoundment to design water levels. The ensuing breach, reaching depths of 50 m with a width of 2.5 mm, triggered significant seepage at approximately 483 L/s. A two-year effort to lower water levels successfully halted hydraulic fracturing [4]. Similarly, Austria’s Kölnbrein arch dam (H = 200 m) faced analogous challenges. Cracks emerged in the dam heel upon its 1978 commissioning, resulting in substantial leakage (200 L/s) at reservoir water levels reaching 158 m. Diverse measures, including freezing, grouting, and innovative anti-seepage interventions, failed to mitigate the issue. As a result, a downstream multifaceted support arch was introduced to enhance stress distribution and restore operational stability [5]. Drawing insights from the Kölnbrein experience, Austria’s Zillergrundl arch dam (182 m high) incorporated a bottom joint and pre-dam heel apron in its design. In addition, post-impoundment, hydraulic fracturing persisted just above the bottom joint. Mitigation necessitated reservoir drainage, transforming the dry bottom joint into a wet joint, coupled with controlled water pressure moderation [6].

The Xiaowan arch dam, towering at 294.5 m, boasts a parabolic double-curvature design and a substantial water thrust of approximately 19 million tons. The reservoir’s formidable water flow, however, poses maintenance challenges due to its river’s robust discharge [7]. The safety evaluation of the Xiaowan high arch dam has been explored through geomechanical model tests and numerical simulations. Model tests effectively replicate water pressure using jacks and facilitate continuous load addition until failure occurs, enabling the attainment of overload safety factors across diverse ultimate states [8]. Yang’s approach [9,10] combines overloading and strength reduction analysis, encompassing a 20% weakening of structural surfaces. Results indicate that dam heel cracks manifest at safety factors ranging from 1.2 to 1.4, with maximum overload safety factors of 3.96 to 4.20. References [11,12] underscore safety factors of 1.2 to 2.5 during upstream surface and dam heel initiation cracking, with overload failure safety factors ranging from 5 to 6.5. Jia’s assessment highlights substantial upstream surface cracking risk for the Xiaowan arch dam [13]. Meanwhile, Zhou and Jin prognosticate the dam’s cracking stability by juxtaposing it with the Ertan and Kölnbrein arch dams listed in

Table 1. The ultimate safety factors of analogy engineering [14,15].

Arch Dam	The Safety Factor of Initiation Cracking in Dam Heel K1	The Safety Factor of Crack Propagation K2	The Safety Factor of Overall Instability K3
Xiaowan	1.1	2.0	3.5–4.0
Ertan	1.3	2.0	4.5–5.0
Kölnbrein (before reinforcement)	0.9	1.65	2.56
Kölnbrein (after reinforcement)	1.2	2.0	3.0

[14,15]. Outcomes reveal a dam heel cracking safety factor of 1.1 for the Xiaowan arch dam, surpassing Kölnbrein’s unreinforced dam yet trailing Ertan and Kölnbrein’s reinforced states.

Diverse strategies were implemented to ensure the safety of the Xiaowan arch dam. These encompassed the utilization of C₁₈₀45 concrete in high tensile stress zones above the dam foundation and the establishment of induced dry joints near the dam heel foundation to optimize stress conditions [16,17]. However, considering the pronounced risk of cracking and its potential to culminate in hydraulic fracturing, the efficacy of these measures in ensuring safety remains insufficient. Consequently, a hydraulic fracturing model test was conducted to investigate the post-crack anti-hydrofracturing capability of concrete. This pursuit aimed to identify the requisite engineering interventions essential for averting hydraulic fracturing within the Xiaowan arch dam.

Leakage issues arising from surface cracks in concrete dams are typically averted through the application of polymer geomembranes or coating materials onto the dam’s surface [18,19,20]. As per ICOLD Bulletin 135, approximately 400 dams worldwide have incorporated PVC geomembranes to mitigate seepage concerns [21]. Coating materials encompass impregnated silicone, fluorocarbon resin, epoxy resin, and polyurea, et al. Silane impregnation treatment on concrete surfaces for global and domestic projects, meticulously elucidating mechanisms, was summarized by Li [22]. Meanwhile, Song examined the utility of fluorocarbon resin coating on the Guandi Hydropower Station dam surface [23]. Additionally, Hao’s work focused on repairing cracks and voids on the Xinanjiang spillway’s concrete surface using epoxy mortar [24].

Innovatively, polyurea emerged as a novel dam surface protection material in China around 2008. Notably, Hao et al. [25] and Jia et al. [26] developed GB plastic sealing material from IGAS plastic filler for CFRD in 1995, exhibiting commendable fluidity, water resistance, and extended freeze-thaw cycles. The GB material, heralded for its robust concrete adhesion, found extensive application in China and abroad, effectively replacing traditional water stops for concrete joints [27,28]. In light of the imperatives of economic safety and temperature-humidity stability near the dam heel, the synergy between polyurea and GB plastic material is being explored as a countermeasure against surface hydro-fracturing for the Xiaowan arch dam.

This study delves into hydraulic fracturing in a 300-m super-high concrete arch dam, specifically focusing on tensile cracks on the dam surface and near the dam heel when subjected to high-pressure water. Using the Xiaowan arch dam as a foundation, a simulation test method for hydraulic fracturing in full-graded concrete structure specimens with cracks under tensile and compressive stress is designed. The results from hydraulic fracturing model tests led to the proposal of a novel dam surface anti-hydrofracturing technology. A pioneering flexible anti-hydrofracturing system, combining GB plastic body and polyurea elastomer, is introduced for super-high arch dams. This system’s reliability is verified using specialized large-scale indoor test equipment, confirming its efficacy under 3 MPa water pressure and 8-millimeter crack openings. These measures significantly mitigate hydraulic fracturing risks and bolster dam safety, offering crucial support for similar endeavors such as the Liyuan concrete face rockfill dam and the Fengning Pumped-storage Power Station. Following this research, recommendations have been adopted by the Xiaowan arch dam, leading to improvements since its impoundment in 2008.

2. Hydraulic Fracturing Test

2.1. Specimen

A cylindrical specimen made of fully graded concrete similar to the concrete used in the Xiaowan arch dam is designed prior to its construction (depicted in Figure 1). The designed dimensions comprise a diameter of 450 mm and a height of 1350 mm. The central portion of the model features a circular crack region with a 150-millimeter diameter. This crack area parallels the specimen’s bottom and follows procedures detailed in Ref. [1]. Concrete materials are outlined in

Table 2. Concrete composition.

Materials for One Cubic Concrete (kg)						Compressive Strength (MPa)
Cement	Fly ash	Water	Sand	Stone	Water reducer	7 d	28 d	90 d	180 d	365 d
147	63	84	533	1688	1.05	28.5	40.3	50.5	55.3	57.6

. A piston hydraulic pump replicates consistent high-water pressure loading, while a 15,000 kN universal testing machine enables the realization of compressive or tensile stress conditions at the crack location section.

2.2. Procedure

Upon mounting the specimen onto the universal testing machine, an initial load of 2 kN is applied for stabilization. Subsequently, the specimen’s inlet is connected to the pump, while the outlet links to both a water pressure meter and a water stop valve. Prior to initiating high water pressure, the water stop valve is opened, gradually introducing water into the preset crack until clear water flow verifies air removal. The valve is then sealed, and the pump is halted. The hydraulic fracturing test commences by applying a constant vertical stress ${\delta }_y$ at the concrete specimen’s ends. The high-pressure water load $p_w$ is then introduced to the predetermined cracks through a stepped loading approach (Figure 2). Each pressure level is maintained for 30 min until specimen failure signifies the onset of hydraulic fracturing. Experimental and numerical findings confirm a normal loading case’s presence of a 1.1 MPa tensile stress at the Xiaowan arch dam’s dam heel [12]. Hence, a vertical tensile stress of 1.1 MPa is emulated in the specimen. Moreover, hydraulic fracturing failure tests were conducted for stress-free and compressive states at −1 MPa to gauge the specimen’s performance under distinct conditions.

2.3. Results and Analysis

Test outcomes reveal distinct characteristics during hydraulic fracturing in concrete specimens. Initially, the cylinder’s central surface becomes damp, followed by the emergence of small droplets. A distinct, boring sound accompanies the appearance of a noticeable surface crack, leading to forceful water expulsion (illustrated in Figure 3). Simultaneously, the pump’s inability to sustain water pressure leads to a rapid pressure decline. Notably, the sound and crack morphology differ under compressive stress compared to tensile stress, exhibiting thinner and shorter crack lines under compression.

Fracturing water pressure, denoting the peak pressure the pump reaches, is recorded as the specimen’s hydraulic fracturing threshold. Results, as tabulated in

Table 3. Summary of hydraulic fracturing under different vertical stresses.

No.	Age (d)	Tensile Strength ${\mathit{f}}_{\mathit{t}}$ (MPa)	Vertical Stress ${\mathit{\sigma }}_{\mathit{y}}$ (MPa)	Fracturing Water Pressure ${\mathit{p}}_{\mathit{w}}$ (MPa)
I-1	7	2.15	0	1.5
I-2	28	3.20	0	2.4
I-3	365	4.32	0	2.7
II	28	3.12	−1.0	3.2
III	90	4.16	1.1	2.3

, underscore the correlation between concrete’s hydraulic fracturing resistance, tensile strength, and stress state. Higher tensile strength correlates with greater hydraulic fracturing resistance. Compressive stress increases cracked concrete’s high-pressure water resistance, while tensile stress diminishes it.

Concrete specimens I-3 and III exhibit fracturing water pressures of 2.7 MPa and 2.3 MPa, respectively, both below the dam heel’s hydrostatic pressure (~3.0 MPa). Consequently, surface cracks near the dam heel entail the risk of hydraulic fracturing. Tensile stress mainly manifests on the upstream surface near the dam heel during standard operations. To ensure safety, an anti-hydrofracturing approach is recommended within the 30 m area above the dam heel to thwart hydraulic fracturing at the Xiaowan arch dam. By maintaining dam section water pressure 30 m above the foundation below 2.7 MPa, safety can be assured based on this analysis.

3. Flexible Anti-Hydrofracturing Solution

An independent and flexible anti-hydrofracturing solution has been introduced, targeting the upstream surface spanning from the foundation up to a point 30 m above it (delineated in Figure 4). The procedure entails affixing GB plastic sealing material onto the concrete surface, followed by a layer of polyurea material sprayed onto the GB material. Notably used in CFRD joints, GB plastic sealing material boasts exceptional adhesive properties and the capacity to accommodate extensive joint deformations. In instances of concrete surface cracks, the GB sealing material functions as mastic, adeptly filling cracks under heightened water pressure. Concurrently, the surface-applied polyurea coating serves as an impervious barrier against hydrofracturing.

3.1. Laboratory Test

3.1.1. Model

Most perilous dam surface cracks tend to arise during operation, necessitating the safety of the flexible anti-hydrofracturing system as cracks widen under high water pressure. To validate this, an anti-hydrofracturing system simulation test model (Figure 5a,b) was devised. The model comprises upper and lower sections connected via steel composite beams and operated using screws. The upper segment, fortified with a 20-millimeter steel flange and welded structural reinforcement, is sealed with a 20-millimeter-thick steel plate top cover fixed by 44 bolts. Water is introduced, and hydraulic pressure is applied using a piston hydraulic pump.

The lower part mirrors the upper, featuring a 20-millimeter steel plate overlay and steel bars atop which a 10-centimeter-thick concrete layer is set on the frame’s inner wall. The concrete surface is secured with GB plastic sealing material and polyurea. Four 100-ton hydraulic jacks are evenly positioned between the top cover and the steel composite beam. During testing, hydraulic jack synergy with water pressure facilitates controlled upper part movement, achieving precise crack opening management.

3.1.2. Materials

Typical surface crack widths in concrete dams range from 1 mm to 2 mm, although larger cracks like the 8-millimeter-wide crack in the Shalaohe concrete arch dam have been observed [29]. Thus, the safety of the authors’ flexible anti-hydrofracturing systems hinges on accommodating an 8-millimeter concrete crack width under 3 MPa water pressure. Consequently, the thickness and elongation properties of the sprayed polyurea are pivotal to managing water pressure.

For GB sealing filler material and sprayed polyurea, static tensile tests were conducted after 7-day curing. The test proceeded at a rate of 500 mm/min, yielding the results shown in

Table 4. The mechanical properties of flexible protective materials.

Material	$\mathbf{Density}/\mathbf{g}·{\mathbf{cm}}^{-3}$	Tensile Strength at Break/MPa	Elongation at Break/%
GB	1.52	-	300
Sprayable polyurea	1.05	22	400

. Polyurea exhibited an average maximum elongation exceeding 400% and an average maximum tensile strength of 22 MPa. Extensometer control governed elongation, while micrometer measurements tracked section thickness at every 50% elongation increment, with their relationship depicted in Figure 6. To ensure sprayed polyurea safety, a post-stretch thickness of not less than 2 mm is required for an 8-millimeter crack opening under 300 m water pressure. Consequently, the sprayed polyurea thickness should not fall below 4 mm.

3.1.3. Schemes

Drawing from the investigation of GB and sprayed polyurea materials, three distinct combinations of these materials were formulated for the model concrete’s interior surface (Figure 7). Employing reserved joints in the middle, the four corners underwent a process: initially, they were lined with 3-millimeter-thick GB material, followed by a polyurea coating of 10 mm in thickness. This measure ensured the preservation of the model’s structural integrity at the corners.

After a one-week polyurea application within the model, water was introduced, and water pressure was gradually raised to 3 MPa. Subsequently, jacks were adjusted to achieve a 5-millimeter model opening sustained for 24 h. In an orderly test, joint expansion reached 8 mm and continued for a week. Test success dictated further joint expansion. Throughout testing, the polyurea’s depth within the concave crack was gauged, and the polyurea coat’s actual thickness was measured post-test completion.

3.1.4. Results and Analysis

The simulation model tests demonstrated no leakage or damage following a sustained water pressure increase of 3 MPa for 24 h. Subsequently, the reserved joint width was incrementally expanded to 5 mm, 8 mm, 10 mm, 15 mm, and 20 mm. Notably, the system encountered failure due to leakage when the reserved joint reached 20-millimeter wide.

Table 5. The elongation under different crack openings.

Crack Width b/mm	The Elongation of Scheme 1	The Elongation of Scheme 2	The Elongation of Scheme 3
5	0.46	0.49	0.35
8	0.90	0.87	0.67
10	1.17	1.21	0.83
15	1.48	1.62	1.12

presents the elongation of the three flexible anti-hydrofracturing systems across varying joint widths. Elongation (ε) was computed using the Giroud Formula (1) [30]. For the anti-hydrofracturing systems under a long, wide slit, the strain was modeled as an arc. Parameter “b” represents the reserved joint width, and “h” stands for the maximum depth into the crack, gauged using a depth micrometer.

$$\epsilon =\left\{\begin{array}{l}\frac{1}{2}\left(\frac{2h}{b}+\frac{b}{2h}\right)\cdot \arcsin {\left[\frac{1}{2}\left(\frac{2h}{b}+\frac{b}{2h}\right)\right]}^{-1}-1\hspace{1em}\hspace{1em}\hspace{1em}h\le \frac{b}{2}\\ \frac{2h}{b}+\frac{\pi }{2}-2\hspace{1em}\hspace{1em}\hspace{1em}h>\frac{b}{2}\end{array}\right.$$

(1)

As evident in Table 5, the three proposed anti-hydrofracturing system choices aptly align with criteria involving concrete crack expansion of up to 8 mm under 3 MPa water pressure while retaining elongations below 200%. The measured sprayed polyurea thickness for scheme 1 is 4.2 mm, scheme 2 is 6.5 mm, and scheme 3 is 8.5 mm. Minimal stretched polyurea thickness stands at 2.8 mm for scheme 1, 2.9 mm for scheme 2, and 3.6 mm for scheme 3. All three systems meet the anti-hydrofracturing system requirements.

3.2. Onsite Scheme Construction

3.2.1. Test

For the field test, the chosen site is the upstream surface of the No. 13 dam section. Test areas A and B, each encompassing 28 m², were designated on this surface (Figure 8a). Test area C was positioned near the corner of the upstream surface and the fillet, as depicted in Figure 8b. Schemes 1 and 2 were assigned to areas A and B, respectively. In consideration of the induced dry joint from the No. 17 to No. 28 dam section, scheme 3 was applied to test area C. This scheme entailed incorporating a GB filler rod (for support), 3-millimeter GB material, a segment devoid of GB lining, and a 7-millimeter-thick layer of sprayed polyurea. Prior to the anti-hydrofracturing treatment, the concrete surface needed to be clean, dry, and free of impurities such as dust and oil. Polyurea could be directly sprayed onto the GB plate.

Over the next fortnight (two weeks), scrutiny aimed to detect any deformation or detachment between the anti-hydrofracturing system and the concrete during construction. Following this period, adhesion strength tests were conducted, and the thickness of sprayed polyurea underwent assessment.

The outcomes of the field test exhibit no signs of deformation or detachment between the anti-hydrofracturing system and the concrete throughout the initial 20-day period following polyurea application. Employing an adhesion strength tester (Figure 9), pull-out failures were assessed across the three test regions. Notably, maximum adhesion strength and locations of tensile failure were noted, along with confirmation of sprayed polyurea thickness. Test results, as presented in

Table 6. Results of adhesion strength.

Test Area	Adhesion Strength /MPa	Measured Thickness/mm	Failure Position
A	0.24	5.67	80% disconnected from the GB board 20% disconnected from standard block and polyurea
B	0.37	6.47	80% disconnected from the GB board 20% disconnected from standard block and polyurea
C	2.31	7.38	100% disconnected from standard block and polyurea

, reveal that the sprayed polyurea thickness in all three test zones surpasses the original design, thus meeting the design criteria. In the A test area, the average adhesion strength between the GB plate and concrete exceeds 0.23 MPa, while in the B test area, the average adhesion strength between GB glue and concrete surpasses 0.37 MPa. In the C test area, failure positions occur at junctions between standard blocks and polyurea, indicating robust concrete-polyurea bonding. The average adhesion strength between polyurea and concrete is consistently above 2.31 MPa.

The aforementioned test outcomes demonstrate the feasibility and suitability of all three schemes for implementation. Among these, scheme 1 stands out with its smooth appearance and enhanced ease of construction quality control, as depicted in Figure 10. Additionally, scheme 1 boasts a thicker, more flexible, and more cost-effective GB plate compared to scheme 2, making it better suited for accommodating significant deformations. Consequently, it is recommended that scheme 1 be prioritized as the primary choice for the official upstream dam surface protection approach, while scheme 3 is advised for application in the corner between the upstream surface and the fillet.

3.2.2. Application

The anti-hydrofracturing system on the Xiaowan arch dam’s upstream surface employs a grid-like protective block configuration (refer to Figure 11a,b). Each block measures 5 m in height and 10 m in length, with a 30-centimeter gap between blocks. Within these blocks, two schemes are employed: Scheme 1 involves applying a 3-millimeter-thick GB plate and a 4-millimeter-thick polyurea spray, while Scheme 3 utilizes direct 7-millimeter polyurea spraying in interval areas to prevent block overlap. SK hand-scraping polyurea, 20 cm in width, is used to edge-seal the sprayed polyurea. At the upstream surface-fillet corner, Scheme 3 combines a 3-millimeter GB plate with a 7-millimeter polyurea spray. Adjacent to the dam heel’s induced joint, 2 m of fly ash backfill is used, transitioning to sand, and then covered with backfilled rock slag for protection. Commencing in December 2007, the flexible anti-hydrofracturing system for the Xiaowan arch dam was completed in May 2009, covering a 19,339 m² area from dam sections No. 7 to No. 36. Leakage detection began in December 2008, revealing a mere maximum leakage of 2.78 L/s post-implementation of the proposed scheme. This minimal leakage stands as the smallest among comparable global projects, such as Jinping I (H = 305 m) at 64.9 L/s, Xiluodu (H = 285.5 m) at 24.9 L/s, Laxiwa (H = 250 m) at 6.0 L/s, and Udonde (H = 270 m) at 6.3 L/s [31], underscoring the efficacy of the authors’ system in meeting expansive underwater sealing needs with expedited construction requirements.

4. Conclusions

In this study, a groundbreaking methodology for evaluating hydraulic fracturing and a cutting-edge proposal for a flexible anti-hydrofracturing system for the Xiaowan arch dam have been presented. Through a comprehensive blend of laboratory tests and pragmatic project integration, these conclusions have been not only established but also fortified:

(a)

The Xiaowan arch dam poses a hydraulic fracturing risk, prompting the proposal of an independent anti-hydrofracturing system for the upstream area, extending 30 m above the foundation.

(b)

Three simulation-tested schemes, namely ‘3 mm GB plate + 4 mm polyurea’, ‘1 mm GB glue + 5 mm polyurea’, and ‘7 mm polyurea’, are suggested. These schemes prove safe for dam surface cracks up to 8 mm wide under 3 MPa water pressure.

(c)

The feasibility of the schemes is assessed through field tests, selecting a scheme involving pasting a 3-millimter GB plate and spraying 4-millimeter polyurea on the upstream surface, along with local 7-millimeter polyurea spraying.

(d)

The recommended anti-hydrofracturing scheme has been applied to a 19,339 m² surface area of the Xiaowan arch dam from No. 7 to No. 36 dam sections. The maximum leakage post-2008 impoundment under normal water levels remains 2.78 L/s, the lowest among similar dams worldwide.