Svoboda | Graniru | BBC Russia | Golosameriki | Facebook
Next Article in Journal
Impact of Long-Term Storage on Mid-Infrared Spectral Patterns of Serum and Synovial Fluid of Dogs with Osteoarthritis
Previous Article in Journal
A Methodology for Identifying Coastal Cultural Heritage Assets Exposed to Future Sea Level Rise Scenarios
Previous Article in Special Issue
Geological Modeling of Shale Oil in Member 7 of the Yanchang Formation, Heshui South Area, Ordos Basin
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Applied Sciences
Volume 14
Issue 16
10.3390/app14167212
applsci-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Exploring the Ground-Penetrating Radar Technique’s Effectiveness in Diagnosing Hydropower Dam Crest Conditions: Insights from Gura Apelor and Herculane Dams, Romania

by
Alexandra Georgiana Gerea
1,2,3,* and
Andrei Emilian Mihai
1
1
National Institute for Earth Physics, 077125 Bucharest, Romania
2
Doctoral School of Geology, University of Bucharest, 010041 Bucharest, Romania
3
School of Electronic, Electrical and Computer Engineering, University of Birmingham, Birmingham B15 2TT, UK
*
Author to whom correspondence should be addressed.
Appl. Sci. 2024, 14(16), 7212; https://doi.org/10.3390/app14167212
Submission received: 8 December 2023 / Revised: 4 August 2024 / Accepted: 13 August 2024 / Published: 16 August 2024
(This article belongs to the Special Issue Advances in Geosciences: Techniques, Applications, and Challenges)
Browse Figures
Versions Notes

Abstract

:
When it comes to hydropower dam safety, continuous and comprehensive monitoring is increasingly important. Especially for aging dams, this can pose a difficult challenge that benefits from a multimethod analysis. Here, we present the use and suitability of a geophysical method, Ground Penetrating Radar (GPR), for the non-invasive assessment of two distinct types of hydropower dams in Romania: Herculane (a concrete arch dam) and Gura Apelor (an embankment dam with a rockfill and clay core). Unlike traditional monitoring methods for dam safety in Romania, which might provide an incomplete overview, GPR offers a broader, non-destructive approach to evaluating some elements of dam integrity. Here, we present the results of surveys carried out with a 200 MHz antenna on the crests of both dams. The aim was to conduct a rapid assessment of the crest condition and identify the potential damage to the crest that may elude standard monitoring techniques. The surveys provide an imaging indicative of the structural integrity, although this is more challenging in the embankment dam, and additionally we provide significant information regarding the deformations in the upper layers. This complements data from routine topo-geodetical surveys, offering a potential explanation for the vertical displacements observed therein. We highlight several areas of potential deformation as well as degradation in subsurface structures such as rebars. The results underscore the value of GPR in supplementing established dam monitoring methods, highlighting its effectiveness in different contexts and dam types, as well as its potential in shaping future standards for dam safety management in Romania.

1. Introduction

Climate change has emerged as one of the most pressing societal challenges of the 21st century, and transitioning to low-carbon energy sources is an essential aspect of tackling this challenge [1]. The deployment of new energy sources is a major lever for decarbonization [2], but established forms of energy, such as existing hydropower dams, are also crucial for this purpose [3].
In Romania, dams have been a core energy supplier for decades. There are currently 545 known hydropower dams in the country [4] that provide over 10% of the country’s primary energy [5] and over a quarter of the country’s electricity [6].
While hydropower dams represent a sensitive subject due to the required tradeoff between clean energy and river conservation, maintaining them represents a very important aspect, not only because they represent massive constructions, but also because of Romania’s commitment to the COP 21 Paris Agreement, where Romania pledged to achieve the 2 degrees target by 2050, and hydropower energy represents a form of green energy that is essential for this goal. This largely involves maintaining existing dams rather than building new ones.
Current legislation does not mandate geophysical surveys as a requirement for tracking displacements, deformations, and degradation, even though the law requires the dams to be regularly monitored using a variety of methods (most significantly, annual topo-geodetical surveys). Therefore, there is a need for testing and finding the best methodologies for applying new methods, especially geophysical methods, which have been shown to provide a variety of useful information which can aid with diagnosing different types of degradation of the dams.
Aging and degradation are well-known problems associated with dams [7,8,9], but early geophysical detection of potential cracks, fissures, and other faults in the dam structure could offer awareness on such potential faults and could direct early intervention that could reduce costs and prevent further degradation. The use of geophysical methods in dam site monitoring must be reliable and robust both in earthfill [10] and in concrete dam sites, as well as both on the dams themselves [11,12] and on the tailings and other areas around the dams [13,14].
However, the GPR method has not been exploited to the fullest in this context. While the literature features significant examples of application of GPR applied in civil engineering to infrastructure like roads, for example [15,16,17,18], the potential of GPR on dams is still insufficiently explored in the context of hydropower dams.
To our knowledge, this is the first time GPR surveys were applied on the Gura Apelor and Herculane dams in Romania, and marks the first attempt to use this method to collect fast data on the crest of the dams, using longitudinal profiles with the purpose of exploring their suitability for the different types of dams normally found in Romania. We aim to provide information which can be integrated with other monitoring methods to provide a fast assessment of any inhomogeneities and deteriorations in the shallow part of the dam’s crest, as well as to explore limitations of the method on these types of dams.

2. Location

The two dams chosen for the study are part of the same hydropower ensemble, owned by Hidroelectrica SA, and are representative for several dams in Romania and across.
The Gura Apelor dam represents the tallest embankment (rockfill with a clay core) dam in Romania and the second largest dam of this type in Europe, after the Thissavros Dam in Greece [19]. It is located in the Hunedoara county (Figure 1) on the south-western edge of the Retezat National Park, approximately 40 km away south from the Hațeg town in the same county. The dam (Figure 1—right) is accessible to the public both by car and foot, with the crest of the dam part of the local road DJ685, popular with both locals and tourists hiking the local mountains. Located at the confluence of the Lăpușnicu and Șes River, tributaries to the Râul Mare river, the dam was commissioned in 1986.
The Herculane dam, on the other hand, represents an arch concrete dam located in the Caraș-Severin county in the Domogled-Valea Cernei National Park, approximately 10 km away north from the famous touristic spa town, Băile-Herculane (Figure 1). Compared to the Gura Apelor dam, the Herculane dam is not open to the public, even though there is a hotel built next to the hydrotechnical facility, and the dam is accessed only by those working at the facility. The dam, commissioned in 1987, is placed on the Cerna river.

2.1. Gura Apelor Dam

With a maximum height of 168 m and a total length of 464 m, the dam’s width at the base is 556 m and at the crest is 12 m with a total volume of 10,285,000 m3 [20]. The foundation rock of the dam (Figure 2) is represented by metamorphic rocks [21], with quartzite schists on the right side and sericeous phyllitic shalles on the left side [22]. The rock material from which the dam was built with is also local and is represented by hard granite schists [20] with a total rock fill volume of 6,422,000 m3. The core (Figure 2) comprises clays with superior plastic qualities, with 50–80% of rock fraction, with a total volume of 1,127,000 m3. This type of clay has a reduced permeability, and it appears to have a self repair property due to these characteristics in case of any fissures [22] and is surrounded by layers of inverted filters with a volume of 183,000 m3. Due to mechanical and morphological asymmetries of the rocks and the valley in which the dam was built, at the base of the dam supporting prisms have been built for the stability of the dam [23], with a volume of 883,000 m3. Other features which are not represented in Figure 2 are the draining gallery, which is found at the basis of the filters on the side facing the downstream, a gallery at the base of the core, with the role of visiting, injecting, and draining, and the spillway and the bottom outlet, which have been placed on right side of the dam. There are actually two spillways built; apart from the bottom outlet at an altitude of 1025 m.a.s.l., the other one is higher at an altitude of 951 m.a.s.l. [24].
The reservoir accumulates a total water volume of 210 × 106 m³, among which 200 × 106 m³ are useful. The accumulation surface is 389.70 ha, with a maximum depth of 162 m. Apart from the hydroenergetic potential of the reservoir, this has another practical use as an anti-flooding mechanism due to its high water capacity potential [25]
The Gura Apelor dam is part of the Retezat hydroenergetic system, with the hydropower plant located underground at approximately 30 m under the thalweg of the Râul Mare river. The plant has an installed power of 335.00 MW and an average energy 560.00 GWh/year [25].
Due to the age and placement of the dam, as this particular dam is located in a narrow valley [22], several refurbishment measures were undertaken in the past few years to make sure that the dam’s life is prolonged for as long as possible. In addition to engineering faults, clay core dams can be affected by earthquakes, faults, or other geological processes [26], and small and large dams alike can manifest settlement and flaking problems [23].
Some crest damage is already visible at the surface (Figure 3—right). However, as is the case with other types of constructions, like bridges or roads, without additional information, it is unclear whether the damage is only superficial or is indicative of more structural problems.

2.2. Herculane Dam

The dam (Figure 1—left) was built on a granitic foundation of the Cerna type, which has a length of approximately 600 m slowly deepening from the upstream toward the downstream under sedimentary deposits [27]. On the left slope of the valley in which the dam is built, at approximately 250 m, the granite is limited by very deteriorated rocks, consisting of clay-shale schists, marl, and marly sandstones [27]. Several measures have been taken to provide stability and consolidate the surrounding area through drilling and injections of concrete [27].
The dam’s structure includes a perimeter gallery for injections, draining, and measuring and controlling equipment, with two access galleries and three stories of gangways on the downstream side (Figure 4). The dam also includes two bottom outlets positioned in the 8 and 9 blocks made from metallic pipes with a diameter of 1200 mm and a total length of 65 m, among which 10.7 m are under pressure, with the spillway positioned under the hydropower plant [27]. Unlike the Gura Apelor dam, where the hydropower plant is situated underground and at a distance from the dam itself, in this case, the hydropower plant is positioned directly at the base of the dam on the downstream side. This setup typifies a surface hydropower plant (Figure 4). It is outfitted with two turbines: one rated at 2 megawatts (MW) and another at 5 MW. These turbines harness water from the Herculane reservoir to generate approximately 12 gigawatt-hours (GWh) of electricity annually [28].
When it comes to the crest of the dam, there are several aspects which can represent minor damages to the dam, as it is the case of the numerous cracks visible from the surface and even patches with grown vegetation (Figure 3—left). From just visual inspections and routine surveys of displacements and deformations, it is difficult to determine whether these cracks are merely superficial damage to the top layer of the concrete or if they extend deeper into the structure. However, both cracks [29] and the presence of vegetation tend to retain moisture in their vicinity. Even if their origin is superficial, this moisture retention can lead to a gradual long-term deterioration of the crest’s surface.

3. Materials and Methods

While several geophysical methods could provide valuable information in this context, here we focus on GPR, and specifically for the assessment of the crest infrastructure elements. Deformations and faults could be detected with the aid of GPR [30,31], especially alongside other geophysical and topogeodetical measurements [32], as they provide subsurface imaging, therefore adding a new perspective and potential understanding for the root cause of some of these issues [33].
GPR is also well suited to the flat continuous surface of dams. As long as good contact between the antenna and the surface on which the surveys are carried out is ensured, the antenna frequency is suitable, and there are no major interferences or sources of noise or attenuation, the GPR method can be used on both concrete and rockfill dams [12,34,35].

3.1. Ground-Penetrating Radar Method

Ground-Penetrating Radar (GPR) is a non-destructive imaging technique used in civil engineering, environmental engineering, and other geophysical applications. It works by emitting electromagnetic (EM) waves into the ground and analyzing the reflected signals to create an image of subsurface features.
GPR is particularly suited for engineering applications as it is non-invasive, fast, and high-resolution. In dams, it can be used for concrete and rebar assessment [7,8,12], embankment investigation [34], leak detection [36], and monitoring of subsidence or other deformations [30,37].
The GPR method uses electro-magnetic signals to detect underground features based on contrasts in the relative permittivity property of the materials. The basic principle behind GPR is straightforward: an antenna with a preestablished central frequency sends electromagnetic pulses into the ground, while another antenna detects the time and shape of the reflected signal. The signal penetrates into the subsurface and then reflects back to the surface; any interference (reflection, refraction, diffraction) creates a detectable signal, and these data can be used to create a profile of the subsurface (Figure 5). An antenna with a determined central frequency will send electromagnetic impulses into the ground, while another antenna will detect the reflections of the signals from the interference of two materials with different relative permittivity values.
A key parameter of GPR antennas is the central frequency; common antenna frequencies typically vary between 50 MHz and 4 Ghz [38]. As it is often the case in geophysics, this parameter generates a trade-off between depth of investigation and resolution: higher frequency offers higher resolution but lower depth of penetration. In the case of studies on dam crests, a low frequency is typically preferable, as it offers an opportunity to inspect potential faults deeper in the structure. Although GPR measurements can also be carried vertically, on the structure of the dam [36], this is challenging logistically, and here only horizontal measurements on the crest of the dam were carried out.
GPR is particularly suitable in engineering applications on paved surfaces because a good contact between the antenna and the surface is essential for high-quality data [38], and paved surfaces can provide good quality contact [38]. Man-made materials are also compact and often exhibit strong relative permittivity contrasts [38], which enables detection. Cracks and fissures are also expected to produce notable detectable contrasts if they are large enough to be detected by the footprint of the antenna.
There are several examples in the literature where GPR is used for comparable applications. Given its high resolution and applicability on paved surfaces, Ground-Penetrating Radar (GPR) is excellently suited for detecting even minute faults in dam elements [34,35,37,39]. In dam monitoring, GPR has already shown its potential to detect voids and cracks inside the dam’s body from several dams located in China [10,11] and areas of internal erosion in dams where the intense leakage was already a known problem on several dams located in Sweden [10], which did not fail due to the self-healing effects of the materials in the filters. However, even though the GPR method has a generally high resolution compared to other geophysical methods, it is very sensitive to site conditions [10] and especially sensitive to clays and areas with high water content which absorbs the radar energy, not allowing for the recording of any useful reflections from underground. Even though it is sensitive to clays, it has been shown that the GPR method could still provide significant information in some cases, described by geometrical changes and horizontal inhomogeneities when applied on embankment dams in Italy affected by surface cracks and deterioration [34]. On concrete dams, GPR proved to be useful in detecting hidden metallic objects (e.g., pipes) in the dams affected by earthquakes in Greece [12] and provide good imaging of the position of other structures found inside the dam’s body, like basements, for example, or galleries. In general, it could be said that, along with other geophysical methods, the GPR method manages to point to areas which require attention [10], as this method still represents one of only methods which could actually scan through the walls or pavements of the dams to provide a subsurface image of any shallow inhomogeneity that might occur inside the dam’s body. On both concrete and embankment dams which have public roads running on top of the crest, GPR can be applied to assess cracks affecting the roads or the quality of the rebars, for example [15,16,17]. The non-invasiveness and short amount of time required for the GPR surveys to take place represent another advantage [35] for this method to be used in dam-monitoring surveys.

3.2. Data Acquisition

All of the data acquisition has been acquired using GSSI SIR 3000 GPR equipment (GSSI Geophysical Survey Systems, Inc., Nashua, NH, USA), a robust and well-established piece of equipment that has been routinely used in the literature [40,41,42]. The antenna used was the 200 MHz antenna on both of the dams. The reason behind the antenna choice was to obtain more information from higher depths and to be able to provide as much information with less signal loss. Even though the antenna does not provide as good resolution as higher frequencies, the objective was to highlight any large-sized variations and trends, as well as providing a more general structural overview of the dam’s crest subsurface.
The surveyed zones were chosen based on their accessibility and on the location’s potential to allow for more information to be obtained with as little interference as possible.
The antenna and the odometer for measuring distance have been calibrated prior to surveying on each particular dam (in accordance with manufacturer recommendations) in order to make sure the recorded distance in the radargrams represents the real distance.
All of the surveys were carried out in dry seasons with no rain forecast before or during surveying.

3.2.1. Gura Apelor Dam—Data Acquisition

There are several key considerations for data acquisition at this dam: the active environment of the dam’s placement, the deformability characteristic to this type of dam, the public road on top of the dam, the visible surface deteriorations, and the subsidence trends highlighted in publicly available leveling data. Given all this, the objective of the surveys carried out on the Gura Apelor dam was to assess and highlight any deformations in the top layers of the dam’s crest (e.g., subsidence and/or uplift) and explore the method’s suitability for detecting discontinuities or inhomogeneities in the first few meters below the crest.
The embankment dam has a clay core, and clay is usually unsuitable for GPR surveys. However, this type of clay (described at Section 2.2) should allow for good signal penetration, particularly due to the small fraction of water from pores, which can be a limiting factor for GPR’s ability to detect features of interest [39].
Therefore, considering all of the above, four profiles (GA1–GA4) were carried out longitudinally on the whole length of the dam’s crest from one end of the dam to the other (Figure 6—left). The profiles were positioned symmetrically from the center of the dam, starting from the access road side towards the spillway gallery. In terms of distance between the profiles, two profiles are located in the middle with a distance between them of 1 m, positioned on each side of the middle line of the dam’s crest (Figure 6), and two other profiles are positioned closer to the edges of the crest (1 m away from the kerb and 3 m away from the closest middle profile).
The purpose of this profile arrangement along the crest was to obtain a general overview of the whole crest and assess the lateral continuity of any features of note. Having the profiles at 1 m away from the lateral sides of the crest also allows for minimizing any interferences in the data resulting from reflections of the nearby structures (fences, raised curb, etc.).
The acquisition parameters (Table A1) included a gain function, negative close to the surface to reduce the effect of the high reflective features which normally appear in GPR data, and 2 more gain points to allow for collecting data with depth. All of the other parameters (range, samples, scan/unit, etc.) were adjusted to allow for signal penetration at a maximum depth possible in the given material (clay) without losing too much information from the surface as well.

3.2.2. Herculane Dam—Data Acquisition

For the Herculane dam, the objective was to determine the type and location of concrete reinforcement, detect potential deterioration in concrete slabs, and conduct an overall concrete condition assessment to explore what other types of features would be visible in the first few meters below the crest. This was particularly relevant as the crest’s surface already presents visible deterioration in the form of multiple cracks and, in some areas, even grown vegetation (Figure 3—left). This vegetation might affect deeper layers by facilitating water infiltrations, thereby deteriorating the concrete reinforcement and slabs.
Similar to the embankment dam in terms of positioning, on the Herculane dam, the profiles were carried out longitudinally from one end to the other of the dam’s crest, and two profiles (H1 and H2) were recorded. Due to the curved shape of the dam, one of the two profiles was positioned in the central part of the crest, with the next one positioned on the side facing the downstream, 1 m away from the first profile (Figure 6—right). Two data sets were collected with the same antenna but different acquisition parameters (Table A1). The first set (S1) was aimed at obtaining more information from the depth, including both low-pass and high-pass filters, as well as a negative gain on the top layer with higher gain applied with depth, and a high stacking number (stacking = 64), whereas the second data set (S2) focused on information closer to the surface and had no filters applied, no stacking, and no gain. The metallic plates visible on the surface (Figure 6—right) were purposefully avoided, therefore ending up with a profile closer to the raised metallic fence.

3.3. Data Processing

The data processing has been carried out using the specialized GPR data processing software, ReflexW (v.9.5), and follows a sequence of filters and gain function applications. Even though the processing flow has some similarities between the data collected on the two dams, the use of filters has been adapted to the characteristics of each dam (Table A2). The rationale behind the steps for data processing was to improve the signal/noise ratio. First, we remove unwanted noise from the data with the application of 1D filters (which act on individual traces) and 2D filters (applied on all the traces), after which the gain functions were applied to amplify and highlight areas of interest; the last step consisted of migrating the data to convert from the time domain to the distance domain, and reducing the aspect of the hyperbolas moving the radargrams closer to the shape of the features in reality. The objective of the data processing steps chosen was to maximize signal/noise ratio to highlight data closer to the central frequency and eliminate information from deeper layers where the data are unclear due to the high attenuation.
The processing flow followed similar steps and functions for the data collected on both dams, where the radargrams recorded a range of 130 ns in the case of the Gura Apelor dam and 180 ns in the case of Herculane dam. The steps are as follows:
(i)
Trace removal using the remove range function, removing the first section from the data (only in the Herculane dam radargrams) covering the data collected over the spillways, which was interfering with the highlighting of other smaller features due to the strong reflection and reverberation effect on the data;
(ii)
Adjusting the zero time with the move startime function;
(iii)
Removing low-frequency noise using the subtract-mean (dewow) function;
(iv)
Removing low and high frequencies outside the range of the antenna frequency using a butterworth bandpass filter;
(v)
Removing system-based coherent noise ringing by subtracting of an average trace with the background removal function;
(vi)
Amplifying signals using the gain function, energy decay;
(vii)
For converting from the time domain to the distance domain, the fk-Stolt migration tool was used (only for the Herculane dam), using a velocity of 0.1 m/ns.
No migration has been carried out on the data from the Gura Apelor dam due to the complex structure and the high variation in velocities throughout the whole radargram. Velocity analysis was carried out using the hyperbola fitting tool in order to approximate velocities in certain areas of interest.
In the case of the Gura Apelor dam, the profiles were approximately 450 m long, which makes visualization difficult for certain features. Therefore, the files were cut on several intervals in order to analyze and highlight features of interest. For the Herculane dam on the other hand, even though the profiles are not as long, at approximately 170 m, the trace removal tool has been used to remove the data collected over the spillway area, which was interfering with other anomalies visible in the data due to their strong visual effect on the data. The remove range tool was also used to extract areas from different intervals to focus on smaller, more specific areas.

4. Results and Discussion

After the data have been processed and analyzed, different areas of interest have been highlighted in the data. Even though all of the data have been processed and interpreted, we present here only some areas of interest. A quantitative analysis was not the aim here. We focus mostly on identifying contrasts indicative of potential deformation, damage, deterioration, or other features of interest for crest assessment [34,35,36,37,38,39].

4.1. The Gura Apelor Data

At the Gura Apelor dam, the collected data yields reliable data up to a depth of approximately 4 m. Although migration was not carried out on this site, this depth estimation is derived from several attempts at fitting hyperbolas, and the images are presented with a depth scale that is calculated based on a velocity of 0.1 m/ns. Below this area, there do not seem to be any significant reflections.
Across all of the profiles, a consistent feature is the presence of three distinct almost uninterrupted horizons (Figure 7 and Figure 8) which span from one side of the dam to the other. In Figure 7, two profiles are shown: profile GA2 (Figure 7—up), which is closer to the middle of the dam, and profile number GA4 (Figure 7—down), which is closer to the edge of the dam facing the reservoir, 4 m apart in between.
The first horizon (Figure 7 and Figure 8) starts at 10 ns on the start of the profile and continues relatively horizontal and uninterrupted with few disturbances.
The second horizon has a different character, as it starts to dip from a time depth about 10 ns to the start of the profile to 20 ns at the 200 m mark, with its lowest point at 25 ns at the 310 m mark, after which it continues in an uptick, ascending back to a time depth of 10 ns with a high point at the 410 m mark, and closer to the right side, it starts slightly to descend again.
The third horizon appears to follow the same pattern as the second horizon. The appearance of the horizons marks a change in permittivity values, probably due to a change in material, and can probably be interpreted as different layers forming part of the road’s foundation on the top of the dam and the sealants on the crest, which is also due to the fact that they appear on all of the four profiles and have a similar character.
Even though these layers appear to be relatively constant in appearance from one side to the other, there are some areas with gaps more visible in Figure 9. These are called gaps just because the reflections are not as intense as the general pattern of the horizon from which they appear to belong. This character might be due to an increase in moisture content, which reduces the speed of the waves traveling through the medium, therefore reducing the intensity of what appears to be a linear anomaly, with higher intensities throughout the profile [43].
Another feature of interest is represented by the area marked by the lowering of the horizons (Figure 7 and Figure 8) with a lowest point at approximately 320 m at 40 ns, which appears to have a distinct character compared to the surrounding area and looks less homogeneous with the third horizon (Figure 8), appearing as more distinct with a stronger reflection in that particular area.
In the same area, there appear to be a multitude of well-contoured hyperbolas, which creates an inhomogeneous aspect, marking a change in subsurface parameters. Compared to the background trend, this area appearing with this combination of characteristics might represent a place of deformation consistent with subsidence.
Another area of interest (Figure 7) appears as inhomogeneous, with the horizon no. 3 slightly higher in intensity, but less intense compared to the area marked as subsidence, with ripples underneath.
Lastly, the area marked as uplift (Figure 7) highlights the area where all of the three horizons have their highest point after the drop appearing in the area marked as subsidence.
Even though all of the features are constant throughout the four profiles at the Gura Apelor dam, it appears that the area marked as subsidence (Figure 7 and Figure 8) has a more distinct character, described as a inhomogeneity, on the middle profiles (GA2 and GA3) compared to the profiles on the extremities (GA1 and GA4).
If these data were to be collected at a small scale, more localized, the general trend in the layers marked with colored lines in Figure 9 would have been easily missed, and, for example, in Figure 8, the layers appear almost constantly parallel inside the area of interest marked a subsidence.
The obtained data are consistent with results from topo-geodetic surveys (Figure 10) published in the literature by Avram et al., 2017 [44]. Specifically, topo-geodetic data obtained with the leveling method is consistent with the subsidence area marked in the GPR data at the Gura Apelor dam. Similarly, there is a good correlation between GPR data and the vertical deformation observed through yearly measurements.
In the topographical data [44], in normal and standard limits, the vertical displacement on the crest of the dam appears to be slightly larger between benchmarks R4 to R7 (Figure 11), with a highest change at the benchmark R5 and R6 located in the same region of the crest where the subsidence is highlighted in the GPR data. For example, according to the provided data (Table A1), the total vertical displacement between surveys carried out in 2016 and 2010 is of about −245 mm at R6 and −240.6 mm at R5; by comparison with the benchmark with the lowest change rate in the same time span, benchmark R12 has a calculated difference of only −39.1 mm.
Noting that the GPR profiles represented start from the side closer to the access road (Figure 10) towards the spillway, marking the start of the radargram closer to R12 and the end closer to R1, therefore, the data in Figure 11 have been flipped to match the true orientation of the GPR profiles radargrams (Figure 11).
This area correlates, also, with the visual observations made prior to surveying, regarding the deteriorations visible at the surface (Figure 3—right), which represents the only place on the whole dam where this type of deformation exists, and its position correlates as well with the area at around 320 m on the GPR profile where the subsidence is marked in the data. In general, it appears that the leveling data pattern throughout the benchmarks from the dam’s crest present a similar trend to the one of the horizons from the GPR data.

4.2. The Herculane Dam Data

Unsurprisingly, the character and the type of anomalies are significantly different at the concrete arch dam, Herculane. After processing the data, it appears that both profiles which were closer to the edge of the dam fail to provide useful reflections, most probably due to the raised kerb and proximity to the fence. Therefore, the focus remains only on the central profile from both data sets.
The first data set, S1 (Figure 12), which has a low-pass and a high-pass filter applied, as well as a negative gain on the top layer with higher gain applied with depth, and a high stacking number (Table A1), appears to offer information from approximately down to 9 m. The velocity for the migration was calculated based on hyperbola fitting, and uses a value of 0.1 m/ns. The profiles have also been cut to extract information only from specific portions of the dam due to the increased number of metallic elements on the surface and the high effect on the data of the spillway area.
Even though a multitude of reflections are visible at the top of the profile (Figure 12—AOI 1), ranging between 10 and 20 ns, these do not appear as strong as the reflections in the second data set (Figure 13—up). The second data set, S2, which has no stacking, no function gain, and no filters applied to the raw data at the acquisition stage, offers a better view at the small continuous reflections in that layer, managing to offer more information regarding the intensity of the reflections as well. For example, in the area marked as AOI 4 (Figure 13—up), these reflections appear less intense, with the AOI 5 slightly more intense compared to the almost non-visible ones in AOI 4. Those marked as AOI 5 (Figure 13—up) represents a distinct area that is almost unique compared to the rest of the profile. In this area, the reflections are probably the strongest. In both data sets, marked as BM with green arrows, significant metallic reflections are represented by reverberation underneath, appearing more intense in the second data set (Figure 13—up).
However, even though the second data set, S2, provides enough detail, the useful information marked by reflections appears to stop at about 90 ns (Figure 13) compared to the data from the first set, S1 (Figure 12); therefore, it does not offer information about the area marked with yellow rectangle named AOI 2 in the first data set (Figure 12). This area is represented by what can be described as an inhomogeneous character compared to the background, with a multitude of reflections with higher intensity.
At the Herculane dam, only the central profile (H1) profile provided useful data, as the fence and raised concrete substantially affected the quality of the data and rendered it largely useless.
The data collected in two different data sets (S1–S2) with two different acquisition parameter settings managed to complement each other into providing more information, compared to using individual data sets, even when using the same antenna frequency.
The second data set at the Herculane dam, S2, manages to image the top layer of the dam, highlighting reflections indicative of rebars, marked as AOI3 at the top (Figure 10). The rebar reflections may provide some information regarding the state of the rebars themselves or their surrounding environment. For instance, the rebars in the area marked as AOI3 (Figure 13—down) show strong reflections that appear indicative of good quality rebars, whereas the rebars in the areas marked as AOI4-5 show a reduced intensity of reflection, which may be indicative of water which lowers the speed of wave propagation [43]. In time, this can lead to corrosion of the rebars. These areas might pose a long-term risk in deterioration at the surface layer of the crest.
The anomalies marked as BM (Figure 13) represent metallic elements due to their very high intensity and reverberation on the vertical; they are most probably the benchmarks used in the topographic leveling surveys which are positioned in the central part of the dam. These elements appear in the first data set as well, but significantly reduced in intensity (Figure 12). As the metallic elements are really just on the surface, these would appear significantly reduced, probably due to the combination of a high stacking number (which indicates the reduced dimension of the metallic elements) and the negative gain function applied at the top of the profile.
The first data set, S1 (Figure 12), may not provide much valuable information in the upper section, but it successfully identifies areas of interest at greater depths compared to the second data set, S2 (Figure 13), where the data becomes less distinct beyond 90 nanoseconds.
The area marked as AOI2 (Figure 12), between 90 and 180 ns, represents a less homogeneous area compared to the surrounding environment, which might represent deterioration inside that area. However, due to the fact that this appears on only one profile, the interpretation of the highlighted anomaly has a high degree of ambiguity. If this anomaly truly exists in reality and is not due to different human errors in data collection or parameters used, it would be located in the blocks 7 and 8 (Figure 4). So far, there is no clear correlation between this behavior and any vertical displacements appearing in the topographic leveling data (non-publicly available), and more profiles and high precision positioning are needed in order to find the relationship between the two.
The surveys may benefit from a more precise positioning system, which has proven challenging here, but this is unlikely to significantly affect the data in this situation.

5. Conclusions

In this paper, we describe two 2D GPR surveys carried out on the Herculane concrete dam and the embankment (rockfill with a clay core) Gura Apelor dam in Romania, marking the first time the GPR method is used on these dams. We attempt to explore the method’s feasibility on both types of dams in Romania by using fast non-invasive longitudinal profiles across the dams’ crest to highlight areas of possible damages or deteriorations and to determine any possible limitations of the method. Although we focus on dams in Romania, the method is applicable to similar environments in other parts of the world as well.
The solid concrete surfaces of both dams ensured consistent close contact between the antenna and the surface. This facilitated excellent signal penetration during the Ground-Penetrating Radar (GPR) surveys and suggests that dam crests are a suitable environment for such surveys.
The GPR technique carried out on the Gura Apelor dam proved to be very useful in providing an overview image of the first 10 m below the crest’s surface, even though embankment dams are usually more challenging in terms of GPR applications due to their high clay content which absorbs the electromagnetic signal.
In this case, choosing the 200 MHz antenna proved to be a good choice, offering a useful compromise between the resolution and depth of penetration. Surveys on the crest of the dam manage to highlight a subsidence area which appears to correlate with the vertical displacements recorded through annual topographical leveling surveys for dam monitoring, enhancing the detail and understanding of the dam’s structure, including aspects not visible on the surface. As well as the subsidence area, the GPR also highlights different layers, which most probably represent the different materials used on top of the crest of the dam, to seal the core and to offer support and foundation for the precast concrete blocks covering the crest to be used as a public road. Even though these are constant across all of the profiles and the area marked as subsidence shows an inhomogeneity character at the lowest point of the horizon no. 1, it is not clear what may be the cause of the subsidence; one set of GPR data collected on a single day cannot provide information about the variation in time.
Therefore, even though these longitudinal profiles on the dam’s crest represent a fast approach to create or provide an overview of the internal structure beneath the dam’s crest, there is a need for such surveys to be carried out regularly and in different environments, as well as integrating the data with other methods (UAV—LiDAR, photogrammetry, other geophysical methods, topo-geodetic displacement surveys, etc.) for increasing the understanding of the development, structure, and causes of this aspect.
The concrete dam, Herculane, highlighted some of the difficulties that can emerge when acquiring GPR data in this context when using lower-frequency antennas like 200 MHz. There is a tradeoff as lower frequency antennas may not provide valuable information regarding minute features such as rebars; however, it was possible here to highlight areas in the crest where the constituent rebars at the surface appear to be affected by humidity (most probably through the surface cracks) which might even be affected by corrosion. This aspect is very important in maintaining the dams in a good state, as untreated degradation at the surface has a tendency to evolve and affect larger surfaces in time, therefore increasing reparation costs. Other possible areas of interest have been highlighted, and a possible inhomogeneity appears at lower levels; however, the main outcome is related to potential corrosions and damages at the near surface related to the cracks visible on the dam’s crest surface. There is, however, a need for more regular profiles to validate the finding as well as more certainty regarding possible damage in the dam’s body related to the inhomogeneity.
A priori knowledge in concrete dams is not a guarantee. Therefore, we must note that while there are similarities between dam crest surveys and other infrastructure surveys (like roads and tunnels), there are also significant differences that must be accounted for at every step, from acquisition to processing and interpretation.
While the 200 MHz antenna was successful in gathering sufficient and varied information for both Gura Apelor and Herculane dams, this does not imply that it will be equally effective for other types of dams; embankment dams are made of local materials, an aspect which implies the variety in terms of permittivity values of the constituent materials, whereas for concrete dams, these may pose different metallic materials on the crest and would interfere with the data collection [45] even more than was the case at the Herculane dam.
It can be said that the GPR method proves to be fast, non-invasive, and effective in providing dam crest conditions and important information about the internal structure of both types of dams built in Romania, concrete arch dams and embankment dams, which can aid at better understanding the behaviors of these types of constructions. We note an area of potential subsidence, also highlighted in previous topo-geodetic surveys, as well as one area of potential rebar damage. We highlight several areas where a more detailed monitoring is suggested and note this as a potential contribution of GPR surveys in this type of environment.

6. Future Work

While we explored the feasibility of this method, more integrated research is required to assess how GPR can be best implemented in long-term dam monitoring. Plans are already being carried out to extend GPR surveys at other hydropower dams in Romania in order to highlight features of interest from inside the dam’s body and possible deformations and damages. Other different methods are being considered for integration with the GPR data as well for a better validation of the results. This includes UAV-based LiDAR and thermal camera surveys for a better mapping of the deformations on the surfaces and to better correlate with the GPR data, as well as aiding with the positioning of these data. In terms of GPR surveys, we will expand surveys to intensify the data collection profiles, creating 3D grids to better image and highlight features of interest. We also consider vertical profiles on the walls of the hydropower dams, only in the cases of the concrete dams (where possible), to correlate with the data from the crest and better image, from multiple directions, the areas of interest and potential deformations inside the dam.

Author Contributions

Conceptualization, A.G.G.; methodology, A.G.G.; software, A.G.G.; formal analysis, A.G.G.; investigation, A.G.G. and A.E.M.; resources, A.G.G. and A.E.M.; writing—original draft preparation, A.G.G. and A.E.M.; writing—review and editing, A.G.G. and A.E.M.; visualization, A.G.G. and A.E.M.; supervision, A.G.G. All authors have read and agreed to the published version of the manuscript.

Funding

This work was carried out within the Program Nucleu SOL4RISC, funded by the Romanian Ministry of Research, Innovation and Digitization, ctr. no.24N/2023, project no. PN23360301.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The datasets presented in this article are not readily available due to the sensitive nature of the constructions and their national importance. The data can be shared, in some contexts, with the approval Hidroelectrica SA which manages the constructions. Requests to access the datasets should be directed to Alexandra Gerea ([email protected]).

Acknowledgments

The authors would like to thank Cornel Paunescu for the help and support throughout the project, to Gerea Gheorghe and Alexandru Micu for logistical support and help with the data acquisition, and to Traian Moldoveanu, Mihai Dutu, and Mihaela Istrate for their help in providing geophysical equipment. Last but not least, we would like to thank Hidroelectrica S.A. for providing access to the dams and to Bijan Nistor for the support.

Conflicts of Interest

The authors declare no conflicts of interest.

Appendix A

Table A1. Data acquisition parameters for both dams.
Table A1. Data acquisition parameters for both dams.
Acquisition ParametersGA1GA2GA3GA4H1-S1H2-S1H1-S2H4-S2
Antenna frequency200 MHz200 MHz200 MHz200 MHz200 MHz200 MHz200 MHz200 MHz
Range150 ns150 ns150 ns150 ns180 ns180 ns180 ns180 ns
Samples5125125125121024102410241024
Format16 bit16 bit16 bit16 bit16 bit16 bit16 bit16 bit
Rate6464646464646464
Scan/Unit1201101109080808080
Gain typeAutoAutoAutoAutoAutoAutoManualManual
Gain points (GP1/GP2/GP3)3 (−20/42/45)3 (−20/42/45)3 (−20/42/45)3 (−20/42/45)3 (−20/45/45)3 (−20/45/45)1 (0/0/0)1 (0/0/0)
High pass filter0000800 MHz800 MHz00
Low pass filter0000100 MHz100 MHz00
Position offset−7.83−7.83−7.83−7.83−12.75−12.75−12.75−12.75
Stacking0000646400

Appendix B

Table A2. Processing parameters, including filters and gain functions of the GPR data. The parameters from up to down are in the order of application.
Table A2. Processing parameters, including filters and gain functions of the GPR data. The parameters from up to down are in the order of application.
Processing StepGura Apelor Profiles (GA1–GA4)Herculane
Profiles Set 1		Herculane Profiles Set 2
Antenna frequency200 MHz200 MHz200 MHz
Remove range (front)-5050
Move start-time−15.104 ns−19.014 ns−19.718 ns
Dewow (time window)5 ns5 ns5 ns
Bandpass butterworth
(lower cutoff/upper cutoff)		70/400 MHz100/400 MHz100/400 MHz
Background removalyesyesyes
Gain function typeenergy decayenergy decaygain function
Gain function parametersscaling value 3scaling value 5scaling value 2.56
FK—migration (Stolt)-0.1 m/ns0.1 m/ns
Velocity analysis
(hyperbola fitting)		0.08–0.1 m/ns0.09–0.1 m/ns0.09–0.1 m/ns

Appendix C

Table A3. Data used to create the graph in Figure 11 based on the data described by Avram et al., 2017 [44].
Table A3. Data used to create the graph in Figure 11 based on the data described by Avram et al., 2017 [44].
Bench-MarkSeptember-2010September-2011June-2012August-2012September-2015August-2016
R1−109.5−118.4−130.1−131.1−169.1−179
R2−176.4−193.9−211.5−215.4−292.9−314
R3−243.2−270.6−294.8−303−411.4−443.8
R4−271.8−304.2−328.1−339.1−458.7−495.2
R5−292.3−329.9−355.3−367.4−493−532.9
R6−288.9−327.9−353.6−367.5−492.2−533.9
R7−253.4−286.1−308.1−319.4−427−461.5
R8−242−273.1−294.3−305.2−407.1−439.3
R9−200.8−225.2−241.7−249.9−328.7−352.2
R10−152.3−166.4−178.8−184.2−235−249.4
R11−112−120.5−129.6−132.5−170.1−179.9
R12−86.5−91.7−97.2−100.9−121.9−125.6

References

  1. Sims, R.E.H. Renewable Energy and Climate Change Mitigation: An Overview of the IPCC Special Report. In Weather Matters for Energy; Troccoli, A., Dubus, L., Haupt, S., Eds.; Springer: New York, NY, USA, 2014. [Google Scholar]
  2. Sinsel, S.R.; Riemke, R.L.; Hoffmann, V.H. Challenges and solution technologies for the integration of variable renewable energy sources—A review. Renew. Energy 2020, 145, 2271–2285. [Google Scholar] [CrossRef]
  3. Siri, R.; Mondal, S.R.; Das, S. Hydropower: A Renewable Energy Resource for Sustainability in Terms of Climate Change and Environmental Protection. In Alternative Energy Resources: The Handbook of Environmental Chemistry; Pathak, P., Srivastava, R.R., Eds.; Springer: Cham, Switzerland, 2020; Volume 99. [Google Scholar]
  4. Costea, G.; Pusch, M.T.; Bănăduc, D.; Cosmoiu, D.; Curtean-Bănăduc, A. A review of hydropower plants in Romania: Distribution, current knowledge, and their effects on fish in headwater streams. Renew. Sustain. Energy Rev. 2021, 145, 111003. [Google Scholar] [CrossRef]
  5. Our World in Data. Available online: https://ourworldindata.org/grapher/hydro-share-energy?tab=chart&country=~ROU (accessed on 7 December 2023).
  6. Our World in Data. Available online: https://ourworldindata.org/electricity-mix#:~:text=Globally%20we%20see%20that%20coal,and%20solar%20are%20growing%20quickly (accessed on 7 December 2023).
  7. Adamo, N.; Al-Ansari, N.; Sissakian, V.; Laue, J.; Knutsson, S. Dam Safety: Technical Problems of Ageing Concrete Dams. J. Earth Sci. Geotech. Eng. 2020, 10, 241–279. [Google Scholar]
  8. Valliappan, S.; Chee, C. Ageing Degradation of Concrete Dams Based on Damage Mechanics Concepts. In Computational Structural Engineering; Yuan, Y., Cui, J., Mang, H.A., Eds.; Springer: Dordrecht, The Netherlands, 2009. [Google Scholar]
  9. Sims, G.P. Ageing of Masonry Dams. Proc. Inst. Civ. Eng. Water Marit. Energy 1994, 106, 61–70. [Google Scholar] [CrossRef]
  10. Adamo, N.; Al-Ansari, N.; Sissakian, V.; Laue, J.; Knutsson, S. Geophysical Methods and their Applications in Dam Safety Monitoring. J. Earth Sci. Geotech. Eng. 2021, 11, 291–345. [Google Scholar] [CrossRef]
  11. Lin, C.-H.; Lin, C.-P.; Hung, Y.-C.; Chung, C.-C.; Wu, P.-L.; Liu, H.-C. Application of geophysical methods in a dam project: Life cycle perspective and Taiwan experience. J. Appl. Geophys. 2018, 158, 82–92. [Google Scholar] [CrossRef]
  12. Karastathis, V.K.; Karmis, P.N.; Drakatos, G.; Stavrakakis, G. Geophysical methods contributing to the testing of concrete dams. Appl. Marathon Dam. J. Appl. Geophys. 2002, 50, 247–260. [Google Scholar] [CrossRef]
  13. Mainali, G. Monitoring of Tailings Dams with Geophysical Methods. Licentiate Thesis, Luleå Tekniska Universitet, Luleå, Sweden, 2006. [Google Scholar]
  14. Kayode, O.T.; Odukoya, A.M.; Adagunodo, T.A.; Adeniji, A.A. Monitoring of seepages around dams using geophysical methods: A brief review. IOP Conf. Ser. Earth Environ. Sci. 2018, 173, 012026. [Google Scholar] [CrossRef]
  15. Rasol, M.; Pais, J.C.; Pérez-Gracia, V.; Solla, M.; Fernandes, F.M.; Fontul, S.; Ayala-Cabrera, D.; Schmidt, F.; Assadollahi, H. GPR monitoring for road transport infrastructure: A systematic review and machine learning insights. Constr. Build. Mater. 2022, 324, 126686. [Google Scholar] [CrossRef]
  16. Batrakov, D.O.; Antyufeyeva, M.S.; Antyufeyev, A.V.; Batrakova, A.G. GPR data processing for evaluation of the subsurface cracks in road pavements. In Proceedings of the 2017 9th International Workshop on Advanced Ground Penetrating Radar (IWAGPR), Edinburgh, UK, 28–30 June 2017; pp. 1–6. [Google Scholar]
  17. Bianchini Ciampoli, L.; Tosti, F.; Economou, N.; Benedetto, F. Signal Processing of GPR Data for Road Surveys. Geosciences 2019, 9, 96. [Google Scholar] [CrossRef]
  18. De Giorgi, L.; Leucci, G. Detection of Hazardous Cavities Below a Road Using Combined Geophysical Methods. Surv. Geophys. 2014, 35, 1003–1021. [Google Scholar] [CrossRef]
  19. Wikipedia. Available online: https://en.wikipedia.org/wiki/List_of_tallest_dams (accessed on 7 December 2023).
  20. Arad, S.; Arad, V.; Veres, J.; Neiconi, T. Stability Monitoring System Implementation for “Rural Mare Retezat” Dam. In Proceedings of the 26th International Symposium on Automation and Robotics in Construction—ISARC 2009, Austin, TX, USA, 24–27 June 2009. [Google Scholar]
  21. Stematiu, D.; Iacob, I.; Sarghiuta, R. Gura Apelor Dam—A Great Achievement of Romanian Dam Engineers. Bul. AGIR 2022, 1, 16–20. [Google Scholar]
  22. Popovici, A. Baraje—Pentru Acumulari de Apa, 2nd ed.; Editura Tehnica: Bucharest, Romania, 2002. [Google Scholar]
  23. Cetin, H.; Laman, M.; Ertunç, A. Settlement and slaking problems in the world’s fourth largest rock-fill dam, the Ataturk Dam in Turkey. Eng. Geol. 2000, 56, 225–242. [Google Scholar] [CrossRef]
  24. Hidroelectrica. Available online: https://www.hidroelectrica.ro/article/78a457a2-376a-849b-c853-88442d7b8c78 (accessed on 7 December 2023).
  25. Hidroconstructia. Available online: https://www.hidroconstructia.com/dyn/2pub/proiecte_det.php?id=44& (accessed on 7 December 2023).
  26. Ohmachi, T.; Tahara, T. Nonlinear Earthquake Response Characteristics of a Central Clay Core Rockfill Dam. Soils Found. 2011, 51, 227–238. [Google Scholar] [CrossRef]
  27. Hidroconstructia. Available online: https://www.hidroconstructia.com/dyn/2pub/proiecte_det.php?id=140&pg=16 (accessed on 7 December 2023).
  28. Hidroelectrica. Available online: https://www.hidroelectrica.ro/article/36 (accessed on 7 December 2023).
  29. Haider, S.W.; Masud, M.M.; Chatti, K. Influence of moisture infiltration on flexible pavement cracking and optimum timing for surface seals. Can. J. Civ. Eng. 2020, 47, 487–497. [Google Scholar] [CrossRef]
  30. Alonso-Díaz, A.; Casado-Rabasco, J.; Solla, M.; Lagüela, S. Using InSAR and GPR Techniques to Detect Subsidence: Application to the Coastal Area of “A Xunqueira” (NW Spain). Remote Sens. 2023, 15, 3729. [Google Scholar] [CrossRef]
  31. Wang, S.; Liu, G.; Jing, G.; Feng, Q.; Liu, H.; Guo, Y. State-of-the-Art Review of Ground Penetrating Radar (GPR) Applications for Railway Ballast Inspection. Sensors 2022, 22, 2450. [Google Scholar] [CrossRef]
  32. Busetti, A.; Calligaris, C.; Forte, E.; Areggi, G.; Mocnik, A.; Zini, L. Non-Invasive Methodological Approach to Detect and Characterize High-Risk Sinkholes in Urban Cover Evaporite Karst: Integrated Reflection Seismics, PS-InSAR, Leveling, 3D-GPR and Ancillary Data. A NE Italian Case Study. Remote Sens. 2020, 12, 3814. [Google Scholar] [CrossRef]
  33. Miskiewicz, M.; Lachowicz, J.; Tysiac, P.; Jaskula, P.; Wilde, K. The application of non-destructive methods in the diagnostics of the approach pavement at the bridges. IOP Conf. Ser. Mater. Sci. Eng. 2018, 356, 012023. [Google Scholar] [CrossRef]
  34. Pueyo Anchuela, Ó.; Frongia, P.; Di Gregorio, F.; Casas Sainz, A.M.; Pocoví Juan, A. Internal characterization of embankment dams using ground penetrating radar (GPR) and thermographic analysis: A case study of the Medau Zirimilis Dam (Sardinia, Italy). Eng. Geol. 2018, 237, 129–139. [Google Scholar] [CrossRef]
  35. Loperte, A.; Bavusi, M.; Cerverizzo, G.; Lapenna, V.; Soldovieri, F. Ground Penetrating Radar in Dam Monitoring: The Test Case of Acerenza (Southern Italy). Int. J. Geophys. 2011, 2011, 654194. [Google Scholar] [CrossRef]
  36. Bigman, D.P.; Day, D.J. Ground penetrating radar inspection of a large concrete spillway: A case-study using SFCW GPR at a hydroelectric dam. Case Stud. Constr. Mater. 2022, 16, e00975. [Google Scholar] [CrossRef]
  37. Rhazi, J.; Dous, O.; Kaveh, S. Detection of Fractures in Concrete by the GPR Technique. In Proceedings of the 16th World Conference on NDT, Montreal, QC, Canada, 30 August–3 September 2004. [Google Scholar]
  38. Reynolds, J.M. An Introduction to Applied and Environmental Geophysics; Wiley-Blackwell: New York, NY, USA, 2011. [Google Scholar]
  39. Xu, X.; Zeng, Q.; Li, D.; Wu, J.; Wu, X.; Shen, J. GPR detection of several common subsurface voids inside dikes and dams. Eng. Geol. 2010, 111, 31–42. [Google Scholar] [CrossRef]
  40. Ristić, A.; Govedarica, M.; Pajewski, L.; Vrtunski, M.; Bugarinović, Ž. Using Ground Penetrating Radar to Reveal Hidden Archaeology: The Case Study of the Württemberg-Stambol Gate in Belgrade (Serbia). Sensors 2020, 20, 607. [Google Scholar] [CrossRef] [PubMed]
  41. Sherrod, L.; Sauck, W.; Simpson, E.; Werkema, D.; Swiontek, J. Case Histories of GPR for Animal Burrows Mapping and Geometry. J. Environ. Eng. Geophys. 2019, 24, 1–17. [Google Scholar] [CrossRef] [PubMed]
  42. Langhammer, L.; Rabenstein, L.; Schmid, L.; Bauder, A.; Grab, M.; Schaer, P.; Maurer, H. Glacier bed surveying with helicopter-borne dual-polarization ground-penetrating radar. J. Glaciol. 2019, 65, 123–135. [Google Scholar] [CrossRef]
  43. Alani, A.M.; Aboutalebi, M.; Kilic, G. Applications of ground penetrating radar (GPR) in bridge deck monitoring and assessment. J. Appl. Geophys. 2013, 97, 45–54. [Google Scholar] [CrossRef]
  44. Avram, D.; Preotescu, C.; Caldarea, G.; Bratosin, I.; Kivu, R. Method of tracking deformations for hydrotechnic dams structures. J. Young Sci. 2017, V, 151–160. [Google Scholar]
  45. Gerea, A.G.; Mihai, A.E. Case Studies and Lessons on Using the GPR Method on Hydropower Dams in Romania. In Proceedings of the 2023 12th International Workshop on Advanced Ground Penetrating Radar (IWAGPR), Lisbon, Portugal, 5–7 July 2023; pp. 1–4. [Google Scholar]
Applsci 14 07212 g001
Figure 1. Location of the two dams. Left—the Herculane (concrete) dam in the Caraș-Severin county. Right—the Gura Apelor (embankment) dam in Hunedoara county.
Figure 1. Location of the two dams. Left—the Herculane (concrete) dam in the Caraș-Severin county. Right—the Gura Apelor (embankment) dam in Hunedoara county.
Applsci 14 07212 g001
Applsci 14 07212 g002
Figure 2. The vertical section of the dam. 1—clay core, 2—inverted filters, 3—rock, 4 and 5—filling prisms. The height of the dam is 168 m and the width of the crest is 12 m. Image reproduced after Popovici, 2022 [23].
Figure 2. The vertical section of the dam. 1—clay core, 2—inverted filters, 3—rock, 4 and 5—filling prisms. The height of the dam is 168 m and the width of the crest is 12 m. Image reproduced after Popovici, 2022 [23].
Applsci 14 07212 g002
Applsci 14 07212 g003
Figure 3. Damage visible at the surface of the dam, on the crest: left—the Herculane dam crest surface; right—the Gura Apelor dam.
Figure 3. Damage visible at the surface of the dam, on the crest: left—the Herculane dam crest surface; right—the Gura Apelor dam.
Applsci 14 07212 g003
Applsci 14 07212 g004
Figure 4. The elements of the Herculane dam visible from the downstream side: dam’s crest, the hydropower plant, the entrance to the gallery, gangways, and spillways.
Figure 4. The elements of the Herculane dam visible from the downstream side: dam’s crest, the hydropower plant, the entrance to the gallery, gangways, and spillways.
Applsci 14 07212 g004
Applsci 14 07212 g005
Figure 5. Representation of the basic operation principles of the GPR data collection and functionality.
Figure 5. Representation of the basic operation principles of the GPR data collection and functionality.
Applsci 14 07212 g005
Applsci 14 07212 g006
Figure 6. Left—the profiles (marked with red) on the Gura Apelor dam (GA1–GA4). Right—the profiles (marked with red) on the Herculane dam (H1 and H2). Not at scale.
Figure 6. Left—the profiles (marked with red) on the Gura Apelor dam (GA1–GA4). Right—the profiles (marked with red) on the Herculane dam (H1 and H2). Not at scale.
Applsci 14 07212 g006
Applsci 14 07212 g007
Figure 7. Processed data from profiles GA2 (up) and GA4 (down) on the Gura Apelor dam. Area of interest marked with rectangles and horizons marked with arrows.
Figure 7. Processed data from profiles GA2 (up) and GA4 (down) on the Gura Apelor dam. Area of interest marked with rectangles and horizons marked with arrows.
Applsci 14 07212 g007
Applsci 14 07212 g008
Figure 8. Close up on the area of interest marked as a subsidence. Processed data from profiles GA2 (up) and GA4 (down) on the Gura Apelor dam.
Figure 8. Close up on the area of interest marked as a subsidence. Processed data from profiles GA2 (up) and GA4 (down) on the Gura Apelor dam.
Applsci 14 07212 g008
Applsci 14 07212 g009
Figure 9. Data from the profiles GA2 and GA4 from the Gura Apelor dam, highlighted with different colors, are the three horizons from the surface.
Figure 9. Data from the profiles GA2 and GA4 from the Gura Apelor dam, highlighted with different colors, are the three horizons from the surface.
Applsci 14 07212 g009
Applsci 14 07212 g010
Figure 10. The topo-geodetic network at the Gura Apelor dam, reproduction after Avram et. al., 2017 [44].
Figure 10. The topo-geodetic network at the Gura Apelor dam, reproduction after Avram et. al., 2017 [44].
Applsci 14 07212 g010
Applsci 14 07212 g011
Figure 11. Leveling data measured at the benchmarks from the dam’s crest, as described by Avram et al., 2017 [44] using data from Table A3. The orientation of the graph matches the orientation of the GPR profiles.
Figure 11. Leveling data measured at the benchmarks from the dam’s crest, as described by Avram et al., 2017 [44] using data from Table A3. The orientation of the graph matches the orientation of the GPR profiles.
Applsci 14 07212 g011
Applsci 14 07212 g012
Figure 12. Processed data from the set 1 (S1). Up—profile H1; down—H1 profile with migration applied.
Figure 12. Processed data from the set 1 (S1). Up—profile H1; down—H1 profile with migration applied.
Applsci 14 07212 g012
Applsci 14 07212 g013
Figure 13. Processed data from the second data set, S2. Up—profile H1; down—profile H1 with migration applied.
Figure 13. Processed data from the second data set, S2. Up—profile H1; down—profile H1 with migration applied.
Applsci 14 07212 g013
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Gerea, A.G.; Mihai, A.E. Exploring the Ground-Penetrating Radar Technique’s Effectiveness in Diagnosing Hydropower Dam Crest Conditions: Insights from Gura Apelor and Herculane Dams, Romania. Appl. Sci. 2024, 14, 7212. https://doi.org/10.3390/app14167212

AMA Style

Gerea AG, Mihai AE. Exploring the Ground-Penetrating Radar Technique’s Effectiveness in Diagnosing Hydropower Dam Crest Conditions: Insights from Gura Apelor and Herculane Dams, Romania. Applied Sciences. 2024; 14(16):7212. https://doi.org/10.3390/app14167212

Chicago/Turabian Style

Gerea, Alexandra Georgiana, and Andrei Emilian Mihai. 2024. "Exploring the Ground-Penetrating Radar Technique’s Effectiveness in Diagnosing Hydropower Dam Crest Conditions: Insights from Gura Apelor and Herculane Dams, Romania" Applied Sciences 14, no. 16: 7212. https://doi.org/10.3390/app14167212

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop