FACULTY OF GEOLOGY AND GEOPHYSICS THE UNIVERSITY OF BUCHAREST ELECTRICAL RESISTIVITY SURVEY FOR UNDERGROUND VOIDS DETECTION IN SALT MINING AREAS… [307110]

FACULTY OF GEOLOGY AND GEOPHYSICS

THE UNIVERSITY OF BUCHAREST

ELECTRICAL RESISTIVITY SURVEY FOR UNDERGROUND VOIDS DETECTION IN SALT MINING AREAS

Scientific Coordinator: Graduate:

Lecturer Dr. Eng. [anonimizat]

2019

FACULTY OF GEOLOGY AND GEOPHYSICS

THE UNIVERSITY OF BUCHAREST

ELECTRICAL RESISTIVITY SURVEY FOR UNDERGROUND VOIDS DETECTION IN SALT MINING AREAS

Case study Ocna Dej and Slanic Prahova

A thesis submitted in fulfillment for the degree of

Master of APPLIED GEOPHYSICS

Scientific Coordinator:

Lecturer Dr. Eng. FLORINA TULUCA

Graduate:

[anonimizat]

2019

Table of Contents

Chapter 1

Introduction

Underground voids detection and mapping is one of the most common aim of geophysical survey. [anonimizat], cellars, caves and others. [anonimizat]-filled or filled with different kind of stuff. Different prospecting techniques have been employed to detect underground voids. Success depends on their ability to reach the target depth with the appropriate resolution for each problem.

Underground voids can be classified by their origin as natural or anthropogenic. [anonimizat], cambering fissures (or gulls), open fault cavities and lava tubes. [anonimizat], [anonimizat], tunnels and shafts.

“Underground voids of natural or anthropogenic origin potentially represent a [anonimizat], [anonimizat]” (Delle Rose et al., 2004; Waltham and Lu, 2007; Parise, 2008, 2010; Zhou and Beck, 2011).

[anonimizat]. [anonimizat], [anonimizat]. Geophysical measurements of the contrasts in physical properties can help define locations of these features. “[anonimizat], or some weak deposit such as mud. They may vary from less than an inch (2-3 cm) in size to many feet (meters), may have any irregular cross section and may be located near the surface or at great depth. The voids may or may not create surface impressions indicating their presence.” (M. C. Hironaka, R. D. Hitchcock, and J. B. Forrest, 1976).

[anonimizat]. One of them is that if we need detailed informations we have to place the boreholes closely. This is a time-[anonimizat]. Also, the interpolation between boreholes can produce significant errors due to the variations of voids. A solution to decrease these drawbacks is to use a geophysical tool.

[anonimizat]-intrusive observations to characterize and map variations in the physical properties of what lies concealed beneath the ground surface. According to the Encyclopedia of Caves and Karstic Science, 2004,“All geophysical techniques require contrasts of some physical properties (density, electrical resistivity, magnetic susceptibility and seismic velocity) between subsurface structures. Although void space in rock may represent an enormous contrast in physical properties that can be advantageous to an investigator seeking concealed caves, underground karst openings are frequently small, irregular targets whose effects are easily masked by those of surface irregularities. In deep exploration, techniques that are useful usually are at the expense of resolution and accuracy; conversely, techniques capable of generating high resolution images of shallow features are often based on high-frequency signals that are rapidly attenuated as they propagate through deeper soil and rock”.

The literature relevant to the identification, mapping or assessment of the subsurface cavities proved that electrical resistivity has a good applicability, which can represents a method to decrease the risks. Also, unknown voids, which can cause hazards or expensive delays to a construction or infrastructure project, can be discovered.

The results of electrical resistivity method are influenced by the geometry, the volume and the depth of underground voids. In Chapter 3, we will see how these results are influenced by the three factors.

Underground voids require permanent monitoring to prevent possible undesirable situations that may occur, which can lead to instability of the area. Instability of the area is caused by landslides, subsidences, salt lake formation, salt dissolution.

Chapter 2

Mining voids & Environmental Problems

Mining operations, specifically high-recovery operations such as longwall and room-and-pillar retreat mines, have the potential to generate overburden deformations propagating from the mined seam to the surface (Peng, 2008). Ground movements induced by underground mining operations may lead to quantifiable impacts on surface and groundwater bodies (Walker, 1988; Matetic, et al., 1995; Singh & Jakeman, 2001; Guo, et al., 2012; Li, et al., 2015). The degree to which mining operations may impact surface and/or groundwater and its recovery cycle is influenced by key mining factors (such as mining depth and coal-seam height, topography, and gob width and length), along with geologic and hydrogeologic parameters (such as percent hardrock, effective porosity, hydraulic conductivity, etc.) and the hydrogeologic conditions of the affected system (such as flow direction and flow quantity) (Agioutantis, et al., 2013; Newman, et al., 2016; Newman, et al., 2017).

There is a need to develop a practical approach to assess potential mining induced impacts in surface and subsurface. This approach should account for key mining factors and key geologic and hydrologic parameters. In the exploitation of salt through dissolution in Romania, the technical rules used worldwide have not always been taken into account. This has led to continued exploitation without taking into account the following:

• Connections between underground voids were not considered to play an important role in the exploitation process, as would be required.

• These connections have led to certain areas, to the partial / total dissolution of the safety pillars and loss of stability of the dissolving ceiling.

• In some cases, the thickness of the projected floor to the formation on the roof of the field was not respected.

• Insufficient knowledge of the existing tectonic context at the level of the formations in the field roof, allowed the design of dissolution voids in unfavorable areas. Their subsequent evolution has generated major imbalance processes resulting in the collapse of the land surface. Such imbalance phenomena can generate a negative impact on the environment due to saturated brine discharges on the surface of the land or in surface emissions.

In Romania the exploitation of salt through dissolution with probes is applied in the case of four deposits of salt, respectively Ocna Mureș, Ocnele Mari, Târgu Ocna and Cacica. The exploitation method by dissolution using probes, forms inside the mountain of salt, systems of various shapes and sizes goals, prompting the change of the magnitude and direction in which the force acts rocks. Exploitation of underground deposit (opening, preparation and abatation) inevitably entails disturbing primary state (natural) voltage rock massif. To ensure the stability of the area, both during the exploitation activity and its cessation, it must be known the direction of action and intensity of deformation processes and the displacement of rocks in mining works, depending on the type of deformation character works (opening preparation, exploitation), the pressure that can develop the outlines of works, displacement and deformation mechanism processes of rocks in space and time. After the extraction of a volume of salt from a reservoir, the state of stresses and strains of the massive salt change, resulting in destruction of surrounding rock stability, fractured rocks on the perimeter of the excavation is put in motion, the movement is transmitting in the massive on a distance that depends on their ability to fill the gap created after the operation. If the hole resulted through exploitation is very high, exceeding the possibilities of the surrounding rock to fill and stop the phenomenon of deformation of rocks, their movement can affect the land surface, causing its degradation and thus destroying the surface or underground targets.

In late December 2010 it took place a phenomenon of subsidence of land adjacent the supermarket "Plus", positioned immediately in south-eastern limit of the salt diapir from Ocna Mureș (Fig. 1). This event was the intrusion of a volume of tailings underground and the expulsion of an equivalent volume of brine, which led to the formation of Lake Plus. Throughout the period of exploitation of salt from Ocna Mures have taken place land surface collapses, which generated at the surface a series of lakes filled with brine (Fig. 2). The peculiarity of the event in December 2010, is that it is the first lake formed outside the contour diapir, all previously formed lakes being placed inside it. At Ocna Mureș, the existence of underground mining voids which are interlinked and collapse of access shafts have created favorable conditions for infiltration of groundwater in underground horizons, causing the weakening of the resistance elements (pillars, floors) and a decline of some works. The effects of this subsidence transmitted to the surface accentuated the depression of the area, which adversely affects groundwater dewatering the river Mureș. The formation of the subsidence funnels and their intersection with groundwater led to the formation of lakes with large dimensions, through which the hydraulic connection is achieved between the phreatic and the backs massive and the flooded underground voids.

Fig. 1 The salty lacustrine complex from Ocna Mureș (Alexe, 2014)

Fig. 2 “Plus” Shop after collapse

The processes associated with this collapsing continues today, extending the crater and the newly created lake surface and extending the fissures around the crater, fissures that that have a the tendency of evolution to several apartment buildings in the area. The phenomena of instability in the field of probes from Ocna Mureș, represent a major risk over the city Ocna Mures, in addition to water and soil pollution by hydrocarbons.

Another aspect is the problem of the historical abandoned mines like Iosif mine from Ocna-Dej mine. Ocna-Dej mine, with a long time activity in salt mining by “dry method”, is another important case of subsidence. “Nowadays, in the investigated zone there are still active mining areas, placed in the vicinity of the old and abandoned ones. Due to the long-time exploitation works which occurred on this site (starting from the Pre-Roman period), the placement of the past-time mining entrances it is not fully revealed yet.” (D. Ioane, F. Chitea, 2015).

The most significantly implications of the presence of salt on the landscape morphology in Ocna Dejului saliferous area are occurrences and development of instability phenomena, both underground and on the surface, in the area of influence of underground voids, resulted from operating activities of salt. Thus, near these operations appeared subsidence depressions. Instability phenomena, widespread in the Ocna Dejului area, are caused, mainly, by high gradient slopes, geological structure, respectively, the presence of impermeable marls and clays, covered with high permeability rocks and the presence of groundwater near surface.

Geophysical methods used to discover the underground cause of the building subsidence were Gravity and Resistivity methods. The results of the data interpretation are illustrated in Figure 3.

Figure 3 Sketch of the geophysical investigated areal – surface view, position of the geoelectrical profiles (D1-D3), flooded mines shafts (M2-M3) and new located bellshape mine entrances using gravimetry (SG) and electrical prospection (SE) investigation methods

( Ioane et al., 2015)

These results shows a good applicability of the mentioned methods. Gravity anomalies is useful for locating underground voids, while status of salt walls, exploitation roof chambers and pillars, highly affected by dissolution processes are obtained by the ERT and SEV sections.

The highest hazard for collapse in the Ocna Dej salt mine was found to be in areas where the walls separating the bell-shaped and rectangular exploitation chambers were already dissoluted, resulting larger underground voids and along the faults affecting the salt body on its entire vertical development, with very active dissolution processes.

Nowadays, Slanic-Prahova is a touristic mine. The mine is closed for extraction purposes and it’s open only for visitors, featuring a microclimate with natural air-conditioning and constant temperature and atmospheric pressure throughout the year.

In the Slănic -Prahova salt mine, the environmental problems started subsequently 1991, when the water amount increased to an alarming level, endangering the safety of the exploitation works. At the ground surface negative landforms of sinkhole type have occurred, while the salt cap rock underwent subsidence that affected the buildings in Slănic town. In order to reduce the hydrogeological hazard to which the mining works are exposed and to stop the progressive deterioration underwent by the buildings situated above or close to the active-karst areas, the infiltration supply area as well as the main flow directions had to be identified, so that a remedial solution was substantiated.

Chapter 3

Study method

Electrical Resistivity Tomography

Tomography means using any kind of penetrating wave for sectional imaging of a structure or object, while the image produced is a tomogram. The method of using electrical resistivity to partition the earth based on varying resistive properties of the earth materials to produce a tomogram is called electrical resistivity tomography (ERT).

Electrical resistivity tomography (ERT) has proven to be an effective geotechnical and environmental engineering tool. Electrical resistivity tomography technique has been successfully used in different situations by numerous investigators (Anderson, et al., 2006; Hiltunen D. R. and Roth M. J. S., 2008; Garman, K. M. and Purcell, S. F., 2008; Loke, M. H., 2008; Zhou, et al., 2002; Zhou, et al., 2000; Hamzah, et al., 2006; Cardimona, S., 2008; Dong, et al., 2008) to assess karst terrains.

In karst terrains, where lateral variations in the depth to bedrock vary greatly, interpolation of the subsurface conditions between two boreholes can often provide erroneous results. The use of electrical resistivity tomography (ERT) can provide more precise information on ground conditions between borehole locations. Like most engineering and geophysical techniques, electrical resistivity tomography (ERT) has its limitations and challenges. For example, if an area is covered by concrete or asphalt, it is difficult to plant the metal stakes used to connect electrodes to the ground for resistivity measurement to be taken. Also, vertical resolution of resistivity data tends to decrease with depth.

Electrical current flow in the subsurface is primarily electrolytic. Electrolytic conduction involves passage of charged particles by means of groundwater. Charged particles move through liquids that infill the interconnected pores of permeable materials (Robinson, 1988). When an electrical resistivity tomography survey is conducted in karst terrain, current flow is generally assumed to be electrolytic rather than electronic.

Variations in the resistivity of subsurface materials are mostly a function of lithology. Information about resistivity variations within the subsurface can be associated with different materials.

Some resistivity values are given below:

Resistivity of common Earth’s materials (Robinson, 1988)

Most materials are characterized by resistivity values that vary by several orders of magnitude. For example, limestone has resistivity values ranging from 50 ohm-m to 107 ohm-m. Most minerals are considered to be insulators or resistive conductors. So in the majority of rocks, electrical current flow is accomplished by passage of ions in pore fluids (electrolytic conduction). The conductivity, which is the inverse of resistivity, is mostly affected by porosity, saturation, salinity, lithology, clay content and to some degree by temperature. Accordingly, materials with constant mineralogical composition can possess different resistivity values, depending on all the above mentioned parameters.

2.1 Principles of the ERT

The ERT was developed at the end of the 1970s (e.g., Edwards, 1977; Barker, 1981; Dahlin, 2001) and rapidly substituted the “classic” one-dimensional geo-electric (Vertical Electrical Sounding and Electrical Profiling) after 2000.

The base principle of the ERT is simple: when a continuous electric current of intensity I (in mA) is generated between two current electrodes A and B – where A is the injection electrode (positive), and B is the reception electrode (negative) – placed at the surface of a ground, which is theoretically considered as homogeneous and isotropic, a semi-spherical electrical field is created (half-space; fig. 4), and its volume is function of the distance between A and B.

Fig. 4 Distribution of the current streams and of the equipotentials in the subsurface and disposition of a quadripole at the ground surface

The more the two electrodes are spaced the more the spatial extent of electrical field is. If we add at this two current electrodes two potential’s electrodes, M and N, allowing the measurement of the difference of potential ΔV (in mV) due to the join action of A and B, the resulting quadripole allows measuring the ground apparent resistivity ρa (in Ωm), which is calculated in the following way (Kunetz, 1996):

ρa = (ΔV / I) k

The geometric factor k (in m) depends on the geometry of the electrode array and the topography, and for each quadripole is:

k = [(AM x AN) / MN] π

2.2 Electrical arrays

For modern electrical resistivity tomography surveys, multi-electrode systems are preferred. The greater the number of electrodes permanently attached to multi-core cable, the higher the investigation capabilities, and less time is spent in the field. Use of multielectrode system allows combination of vertical sounding and horizontal profiling data to be collected simultaneously. Also, it allows the generation of a two-dimensional model of resistivity distribution (Lateral and Vertical).

For 2-D imaging, using a modern multi-electrode system, the spacing between electrodes stays fixed for the entire survey. Measurements are taken sequentially using different sets of four electrodes controlled by switching device. The depth of investigation is a function of the array type, the length of array and the physical parameters of material underlying the area of interest, and typically ranges from one-third to one- fifth of the length of the entire array (Robinson et al., 1988).

The geometry of an electrode array depends on the target depth, the time allowable for data acquisition and the required spatial resolution. When a multi-electrode system is used, the spacing between all electrodes remains the same, while the distance between current and potential electrodes depends on electrode configuration. This distance is controlled automatically by resistivity meter.

Most electrical resistivity tomography surveying is done with one of the electrode geometries illustrated in Fig. 10. For the 2-D Wenner Array (Fig. 10), current and potential electrodes are separated by equal distance ‘a’ such that:

AM = MN = NB = a

All the electrodes are arranged along a continuous line, also known as survey line or traverse. The geometric factor for Wenner Array can be expressed as:

KW = 2 x π x a

For the 2-D Schlumberger array (Fig. 10), the current electrodes A and B are located on the opposite sides from center point of the array .The passive electrodes N and M are placed between A and B electrodes. Suppose the current electrodes A and B are separated by distance ‘L ‘, and the passive electrodes N and M are separated from the center by distance ‘b’, the geometric factor for Schlumberger array can be given by the expression:

KS = π (L2 – b2 ) / 2b

The third geometry is attributed to the Dipole – dipole configuration, where the potential electrodes M and N are not placed between the current electrodes A and B (Fig. 5).

The Dipole – dipole array is logistically the most convenient array used in the field, especially for large scale projects. In this type of array, all four electrodes are placed along the same line, and the distance between the current electrodes A and B is equal to the distance between the potential electrodes M and N, represented by ‘a’, given by the following:

AB = MN = a

The distance between the middle points of current and the passive electrode sets is an integer multiple of a, and the factor itself is assigned to be equal to n (Fig. 5).

The geometric factor K can be found from the following expression:

K DD = π x n (n2 -1) x a

Fig. 5 The most common electrode array configurations

2.3 ERT data acquisition

The data acquisition allows obtaining a pseudo-section in apparent resistivity. Along the acquisition line, the measurement of the apparent resistivity is automatically repeated for every possible combination of the quadripole for the selected array.

The development of electrical tomography techniques has allowed the design of new measuring arrays based on 8 channels equipment. These devices are capable of recording 4 measurements of potential difference in one current injection. The arrays applied is the gradient array, which uses an ‘alfa’ type electrode arrangement where the potential electrodes are between the current ones (ABEM, 2006; Loke, 2014).

A measurement protocol in electrical tomography is the programming and command definition measured by electrode (Fig. 6). For this work it has used the protocol "4-channel multiple gradient" that takes the measures in two cycles called GRAD4LX8 and GRAD4S8. The four measurement channels increase the data acquisition rate, which is faster than single channel equipment. This protocol begins with a long cycle (GRAD4LX8) and at the end of this, the measurement is performed with a short cycle (GRAD4S8). In the long cycle steps, the measurements are conducted between the electrodes of four cables, and in the short cycle only the two central cables are used.

Fig. 6 Schematic diagram of tomography data acquisition

In an ERT survey, the order of electrode spacing affects the lateral resolution of the 2-D ERT profile. As distance between the electrodes increases, the depth of investigation increases, but lateral resolution decreases. During data processing, as an example, when using a 5 ft electrode spacing, the subsurface is divided into rectangular areas that are 5 ft wide with an apparent resistivity assigned to each. By doubling the spacing to 10 ft, the rectangular areas are essentially quadrupled, hence the decrease in lateral resolution on the 2-D ERT profile.

The completed ERT array layout is shown in Figure 7. The resistivity meter (SuperSting) controls the electrodes (e.g., injecting current and measuring voltage) by attaching to two switch boxes (SB1 and SB2) that switch four cables each. Cables 1 to 4 are switched by SB1 while cables 4 to 8 are switched by SB2. The electrodes, resistivity meter, batteries, switch boxes, cables should always be protected from rain and moisture. When inserting the electrodes into the ground, care is to ensure good contact is established between the electrodes and the soil. Ensuring proper contact is especially important during drier conditions when holes formed by the electrodes may not have naturally healed between acquisition days. In addition, in areas where water bodies such as creeks and ditches are present, plastic covers (e.g., Ziplocs) are applied onto the equipment (usually the cable connectors) that contact with water to prevent moisture damage. In areas where a more significant water body is present, ERT data are usually acquired using specially designed marine cables.

Fig. 7 ERT cable and equipment layout DE SCHIMBAT CU ECHIPAMENT MAE

Several factors negatively affect the quality of the acquired ERT data. The most common issue is the unsatisfactory connection/conduction between the stakes and ground. Sometimes, it is difficult to firmly drive stakes into the ground when the ground is hard and dry. Another issues and solutions are illustrated in Fig. 8.

Fig. 8 Common issues and solutions

2.4 ERT data processing

The use of electrical tomography provides a resistivity results sometimes difficult to interpret. For their analysis is necessary to obtain field information and response of the materials present as much as possible. Unfortunately, the results do not always match the expected response due to multiple factors. Forward Modelling is used to approximate the obtained result to the subsurface materials (Loke, 2002). A theoretical profile with known resistivity bodies and simple morphology is modelled to analyse the electrical response. To make these models has been used RES2DMOD software, from Geotomo Software.

This program calculates the apparent resistivity pseudosections for a defined synthetic model. For these apparent resistivity pseudosections generated, the real resistivity is calculated by RES2DINV v.3.59 software, like a field measured profile. To calculate the real resistivity are used the same inversion parameters for the profiles measured in the field. In addition to this posteriori analysis for the materials composing the resistivity profile, has made a priori analysis to define which measuring array should be used for best results.

The resistivity data sets collected in the field were converted into resistivity models for interpretation of subsurface conditions using the RES2DINV software.

ERT data was processed using the following steps:

Inspection of the resistivity data sets for presence of unreasonably high and low negative) resistivity values called “ bad data points” (Loke, 2004).

Removal of “bad data points”.

Compilation of a resistivity model/ERT resistivity profile that displays horizontal and vertical resistivity distribution.

Before processing, the data acquired had to be inspected for presence of “bad data points” (Loke, 2004). “Bad data points” mean resistivities of unrealistically high or low (negative) values. “Bad data points” can be caused by several factors, such as failure during survey of equipment used, for example electrode malfunction. Also, very poor electrode – ground contact can result to “bad data points”. In addition, when a metal stake attached to an electrode is driven into an ice lens, resistivity measurements are affected. Ice acts as an insulator, and affects resistivity measurements. This is a problem for surveys done in winter.

Inspection of “bad data points” is done by viewing a profile plot, illustrated in Figure 9 .The “bad data points” appear as stand out points. All “bad data points” are marked as red plus signs.

The RES2DINV software offers an option that allows for removal of such points manually by simply clicking on them. After the resistivity data sets acquired in the field were inspected and all unrealistic values removed, the RES2DINV software used an inversion algorithm to convert the measured resistivity model/ERT resistivity profiles to a geologic model which reflect lateral and vertical resistivity distribution.

Fig. 9. Example of a data set with a few bad data points

The software creates a resistivity model/resistivity profile that has the same resistivity distribution as the actual resistivity distribution below the corresponding traverse. To increase the quality of the calculated model, the Root Mean Square (RMS) method is used (Loke, 2004). In this method, the smaller the RMS value, the better the calculated model correlates with real resistivity distribution. In this project, an RMS value of 50% was used. To create a resistivity model, the RES2DINV subdivides the subsurface into a finite number of rectangular pixels. Each pixel is assigned a resistivity value which represents the resistivity of different materials encompassed within that discrete pixel; therefore some lateral and vertical smoothing takes place (Anderson, 2006).
The size of the pixels is affected by the spacing between the adjacent electrodes. Horizontal dimension of a pixel is equal to lateral distance between adjacent electrodes, and at shallow depth, the vertical dimension is approximately equal to 20% of the spacing between two adjacent electrodes. With increasing depth of investigation, the vertical dimension of pixels gradually increases up to 100% of the distance between adjacent electrodes (Anderson et al., 2006). The resolution of the output model is a function of the pixel size (Fig. 10). Thus, with increasing depth of investigation, resolution decreases.

Fig. 10. Arrangement of the blocks used in a model together with the data points

When a Dipole-dipole array is used, the maximum depth of investigation is approximately 20%-25% of the array length but this is affected by subsurface condition such as the top layer of the ground being very dry.

The ERT resistivity profiles generated are later interpreted by picking the inverse model resistivity sections. Using the software package RES2DINV, the program generates a two-dimensional model of subsurface resistivity by minimizing the root-mean-square (RMS) difference between the predicted apparent electrical resistivity from an inversion model and the values measured in the field. The model is composed of layers of rectangular model blocks each possessing an electrical resistivity value.

RES2DINV’s time-lapse inversion technique may be used as opposed to inverting each data set independently. Time-lapse inversion allows for the inversion process to begin with previous inversion results as a reference model that assists in constraining later time-lapse data sets (Loke, 1999).

The inversion of a data set for any given date uses the inversion result from the preceding date as the reference model with constraining accomplished using a leastsquares smoothness technique (Geotomo, 2010). Data sets are inverted “simultaneously”, meaning all data sets complete the same round of inversion iteration before starting subsequent iterations. Miller et al. (2008) concluded that simultaneous time-lapse inversion yielded less noisy results compared to other common time-lapse inversion techniques. The ultimate result of the time-lapse inversion method is a more reliable inverted image of the subsurface that not only has minimized the difference between observed and calculated resistivity values, but also preserves the consistency of subsurface features with time.

Caption A (Figure 11) is the measured apparent resistivity pseudosection which represents the data set acquired in the field, caption B is the calculated apparent resistivity pseudosection which represents a synthetic model that is used to estimate the size of the pixels at different layers and C represents the Inverse model resistivity section which represents the true geologic model of the subsurface.

Fig. 11 ERT profiles: A- Measured apparent resistivity B- Calculated apparent resistivity C- Inverse model resistivity

Resolution is a function of electrode spacing and resistivity contrast between lithological different earth materials. The resolution of electrical resistivity tomography (ERT) profile defines the accuracy of interpretation of subsurface conditions. The size of a pixel is a main estimate of ERT imaging resolution. With increasing depth, the vertical dimension of the pixels becomes greater and that reduces ERT resolution. To estimate the size of all detectable objects at a certain depth, it is recommended to compile a synthetic resistivity model. The model can be used to visually estimate the size of the pixels at different depth layers. During ERT survey, when current is induced to flow through deeper layers, the distance between current and potential electrodes is gradually increased. This affects the sensitivity of the ERT method. Gradually increasing the distance between electrodes lowers the intensity of current flow, and accordingly the sensitivity of ERT survey. Thus, interpretation of smaller scale objects at greater depths becomes increasingly difficult and sometimes small objects can be missed or misinterpreted.

Resistivity contrast is another parameter that defines the resolution of ERT profile. When lithological different materials exhibit similar conductivity parameters, sometimes it is difficult to differentiate them simply on the basis of their resistivity parameters. For example, both intact bedrock and air-filled voids typically are characterized by high resistivity values. When an air-filled void is embedded in intact limestone, it typically cannot be easily detected on resistivity profile because of low resistivity contrast.

Chapter 4

Underground Voids Investigation by ERT Method in Salt Mines environment

4.1 Types of Mining Voids

Each form of mining is associated with the occurrence of negative effects in the natural environment. The types of such effects are characteristic for particular mining methods, and the effects are different in the cases of open-pit or underground operations, or solution mining which is also a form of underground mining.

Salt mining is based on underground extraction, the methods being of two types: dry and wet processes. The dry method consists in salt rock cutting by blasting or mining tools and machines. Several extraction systems have been designed under that method (e.g. long wall caving, or chamber mining). The wet (or leaching) methods are based on dissolving salt deposits with fresh water. Brine obtained by the process is subjected to evaporation to regain salt. Wet methods are applied in three options. The first one consists in sprinkling the salt deposit in the virgin salt rock with a nozzle system. That process allows to form workings with specific dimensions and shapes by proper nozzle operation. In that case, the workings system is apparently similar to that obtained by dry methods. The second option relies on dissolving salt rock in the so-called leaching plants or chambers, where salt is dissolved in stagnant water. After specific density has been reached, brine is pumped out and the chamber filled with fresh water. The third option, called solution mining, starts with drilling boreholes from either land surface or underground workings towards salt deposits, followed by mounting pipes in the boreholes. Pipelines are designed to inject fresh water under pressure and collect brine, with various salt saturation. Fully saturated brine is transported for further processing, while unsaturated brine is subjected to a process designed for obtaining full saturation. These processes depend on physical properties of salt rocks which are quite different than those of other types of rocks.

Salt mining method extraction interferes with the original mass rock equilibrium, generating a secondary state of equilibrium that leads to the changes in the state of stress and deformation on workings’ boundaries. Consequently, the rock mass displays a tendency to relocate towards the workings (free space underground). This is the cause that generates processes that lead to the occurrence of underground voids.

The underground voids can be positioned in the salt layer, in the cover formations or hidden by cemented platforms. The position of the underground voids may be influenced by the type of mining (dry or wet processes). For example, through dry method extraction, there may be more problems because the depths are smaller. In deep mining can endanger the stability of the land by the accidental union of the leaching chambers.

Fig. 12 Underground void resulted through the accidental

union of the leaching chambers

Regarding the type, the underground voids can be:

Unfilled / air-filled

Filled with water

Filled with homogenous/relatively homogenous solid material (Backfilled voids)

Therefore, electrical response of the underground voids is different. The rock materials themselves have very high resistivities (hundreds or thousands of ohm meters) and may be considered as mainly a matrix or framework for the contained free water. It is this water and its dissolved salts that allow conduction of the electrical current for measurements. Air also has an extremely high resistivity and may be considered as an insulator at the voltage levels used for the tests. While the effect of large discrete voids is readily apparent, small cracks or voids, when occurring in extreme large numbers, may easily show the same overall effect. In a mining environment, these can be micro-fractures or delamination phenomena above the void which may migrate upward as subsidence phenomena. These cracks may be either air or water filled and will normally be interpreted as void areas.

4.2 Electrical Models of Mining Voids

4.2.1 Synthetic Models

As observed from real field datasets, the electrical response of the underground voids may be quite different. Therefore we tested the ERT response over several types of voids models to extract the most important factors that influences the results in terms of electrical values and shape of the anomaly. Comparing the behavior of different matrix configurations, when the combination of dipole length and dipole spacing varies, is crucial for choosing array matrix configurations to provide greater resolution and good reliability.

An estimative response for ERT results when applied for underground voids detection can be obtained by using numerical models and adequate software programs for solving the forward modelling problem. Therefore, several synthetic models were built, aiming to assess which array type is better for depicting a buried void, either electrically resistive (dry/air filled scenarios) or electrically conductive (water filled or saturated sediments content). For this objective, simplified models of voids placed in a homogeneous medium were considered and estimated results in terms of apparent resistivity sections as well in inversed resistivity sections were obtained using Res2Mode, Res2Dinv and EarthImager software.

First models considered an homogeneous background of 100 Ωm and a high contrasting body, simulating water filled void (Figure 13) and unfilled (Figure 14), with lateral extent of 2m and height 1.2m. Several simulations were made using the EarthImage2D software to study the changes in resolution and reliability when the array type changes.

Successive change of the array type for a conductive and resistive underground void, revealed the complexity of the results, which can be noticed in Figure 13 and Figure 14. The comparison has been shown to be very useful in assessing the efficacy of each used array for such type of survey.

In Figure 13 are illustrated the results obtained using different array types: Schlumberger (A), Dipole-dipole B –, C – Dipole-gradient. The resistivity of the void was kept at a constant value (10Ωm) and fixed geometry for all the trials, in order to observe how the resistivity values and resulted anomaly shape vary by each section. In the case of Schlumberger array (Figure 13a), the resistivity values ranges between 78 – 103Ωm. In the case of Dipole-dipole array (Figure 13b), the range of values is slightly increased, from 62 to 108 Ωm while for the Dipole-Gradient array (Figure 13c), the resistivity values vary from 59.4 to 106 Ωm.

a

b

c

Fig. 13 ERT sections obtained using different electrode array for

a conductive underground void;

(Schlumberger – a, Dipole-dipole – b, Dipole- gradient – c)

For the situations of the unfilled void, with a given value of 107Ωm (practically simulating the infinite resistivity), the results showed that the range of resistivity values do not change much for the used array types: for Schlumberger array, 98-119 Ωm (Figure 14a), for Dipole-dipole, 96-117 Ωm (Figure 14b), for Dipole-gradient, 95-140 Ωm (Figure 14c).

In pseudosections, the geometry of the anomalies is deformed while the inversed sections reveal satisfying results. If it were to extract the geometry, the results are similar. It can be observed that Schlumberger array is predisposed to the occurance of false voids.

a

b

c

Fig. 14 ERT sections obtained using different electrode array for a resistive underground void

(Schlumberger array – a, Dipole-dipole – b, Dipole-gradient – c)

The results revealed that the dipole-gradient and dipole-dipole array configuration offers good results, the anomaly being well contoured and its depth corresponding to the depth of the underground void. This is possible due to the fact that this array type has a great laterally covered area, is sensitive when a lateral resistivity variation is encountered (Chitea & Georgescu, 2015).

In order to observe the influence of the void geometry on the ERT results, synthetic models were further constructed using EarthImager2D/Res2Dinv, to compare resistive underground voids with different lateral and horizontal development while the medium resistivity is constant (100Ωm) using dipole-dipole array type.

a

b

c

d

Fig. 15 ERT sections highlighting the influence of the geometry void

The first section (Figure 15a) presents the results for a small void of 107Ωm, with a lateral extent of 2m, embedded in a layer of 100Ωm, placed at 3.3m. The high resistivity anomaly given by the void, is of a circular shape and well defined (Figure 15a).

The next presented models (Figure 15b-d) are following the effect of two resistive voids of different geometries when placed at variable depths. The first void is placed at 2.8m depth, having 3 x 1m dimensions. The second one is in vertical position, 2.8 – 3.78 m high and 1 m lateral extent. It can be seen that both anomalies have high values, but the anomaly that corresponds to the void in horizontal position is more intense, has higher values. For the second void, it cannot be anticipating the vertical development. The same idea is pointed out from Figure 15c. In this case, the depth of the void with vertical development is increased and the resistivity values of the anomaly are higher than in Figure 15b. The results are influenced by the increasing of the volume of the underground void.

The results revealed that the geometry of the void influences the values of anomalies according to the volume, position and depth of the underground voids. It is an important aspect that will also be highlighted in two-layered models.

Regarding the type of the underground void that can be observed in an area of salt mine works there are several situations which gives different results when investigated by means of ERT method (Figures 16-35). The buildup models are based on two layered geological environment, situation encountered in the real cases studies of Ocna Dej and Slanic Prahova Salt Mines presented in chapter 4.2.2. The top layer was chosen to be very conductive, while the lower layer was given high resistivity values, simulating the salt layer. Salt is known to be very resistive (Watanabe, 2002), but as seen in geoelectrical investigations from the surface, the effect of the conductive strata is highly influencing its value (Kižlo, 2009).

Several synthetic models were constructed using Res2Mod and then, the resulted synthetic dataset was inverted using both Res2Dinv and EarthImager2D software. The two-layered synthetic models are based on the geological setting encountered in real situations at Ocna Dej and Slanic Prahova salt mines, where the resistive salt lens are covered by formations characterized by low resistivity, which seldom drops below 1Ωm, due to the salt enriched fluids that circulates within the vadose zone.

For the undisturbed formations the expected result is illustrated in Figure 16. The anomaly has a linear trend, near the surface are the minimum values (1 Ωm), which correspond to the conductive layer, at the limit between conductive layer and resistive layer is the transition limit, with values between 1.3-2.3 Ωm and below it the resistivity values increase up to 3.0 Ωm.

Fig. 16 Synthetic model 1 – two layered geological environment

The second model (Figure 17) reveals the results in the case of a resistive underground void (104 Ωm) with lateral extent of 9m. The anomaly generated by the void has maximum values of only 3.7 Ωm while the resistivity of the background remained similar to the imposed ones (0.79-1 Ωm). The shape of the anomaly is circular, at its center being the maximum values (3-3.7 Ωm).

Fig. 17 Synthetic model 2

Comparing with the previous, the third situation (Figure 18) shows that if it increases the depth of the same underground void, the anomaly is no longer individualized, it has not its own form, it appears as a deformation, with maximum values by 2.8Ωm.

Fig. 18 Synthetic model 3

If the underground void is at the same depth, the resistivity has the same values by 104 Ωm, but its volume increases, having a vertical extent by 4m, the anomaly is seen as a deformation with a trend towards the surface (Figure 19) and the range of the resistivity values do not change.

Fig. 19 Synthetic model 4

If the underground void is placed at a medium depth (6.7 m), in horizontal position, with a lateral extent by 9 m and vertical extent by 2 m, it can be observed that the high values anomaly tends to individualize (Figure 20). The shape of the anomaly is oval, with values between 1.4-2.7 Ωm.

Fig. 20 Synthetic model 5

In Figure 21 the effect of the underground voids can be compared depending on its geometry. In this way, it can be observed that the effect of the void in horizontal position (9 m lateral extent, 2 m depth) has higher values that the void in vertical position (2 m lateral extent, 6 m vertical extent), although the both voids are at the same depth (Figure 21).

The effects are similar if the underground void in vertical position increases the lateral extent, from 1.5m to 3 m (Figure 22), but if there would be analyzed only the results, the position of the vertical void could not be anticipating, because the resistivity values are in the same range.

Fig. 21 Synthetic model 6

Fig. 22 Synthetic model 7

There were analyzed the effects of two voids. In Figures 23, 24, 25 can be compared the results of three synthetic models changing successive the geometry of three underground voids. Two underground voids remain constant, with the same lateral and vertical extent, only the third void changes its geometry.

In the first situation, the void has a lateral extent of 3 m and 4 m depth, the maximum resistivity value is by 3.5 Ωm. The anomalies can be observed as being well visible and contoured. The second situation (Figure 24), when the void is placed at the limit between the resistive layer and the conductive layer by increasing its depth to 6 m, pointed out that the anomaly for that void is no longer individualized and it becomes as a deformation. In this case, the effect of the other two voids is very little influenced by the third void, the highest values being from 3 to 5.1 Ωm. The third situation (Figure 25), when the void penetrates in the resistive layer, it causes a prominent deformation of the anomaly. Also, the effect of the void in horizontal position is well defined and individualized. The resistivity values vary from 0.66 to 5.2 Ωm.

Fig. 23 Synthetic model 7

Fig. 24 Synthetic model 8

Fig. 25 Synthetic model 9

In Figure 26 it is compared the effects of a vertical underground void, 3 x 5.7 m dimensions, placed in the conductive top layer with the effect of a horizontal void, with a bigger lateral extent (6 m), placed in the resistive layer. It can be noticed that the maximum anomaly is due to the void situated in the conductiuve layer, because the contrast is stronger and it is placed closer to the surface, while the effect of the void from the resistive layer is attenuated.

Fig. 26 Synthetic model 10

For the undisturbed formations the expected results, with higher resistivity values for the bottom layer (104 Ωm), are presented in Figure 25. The effect is similar with the one from the Figure 16, the resistivity values are distributed linear, close to the surface being the minimum values (1 – 1.3 Ωm), increasing with the depth to 3.0 Ωm.

Fig. 25 Synthetic model 11

If it is considered the development of an underground void near the surface, at 2 – 3 m depth, with 6.3 x 2.8 m dimensions, the anomaly is very intense, well defined, having values from 2.7 to 3.8 Ωm (Figure 26) .

Fig. 26 Synthetic model 12

The anomaly is no longer well individualized if the depth of the void increases to 9 m (Figure 27). It became like a deformation, at higher depths than in Model 12 due to the depth of the void. The maximum resistivity value is 3.2 Ωm while in the previous situation 3.8 Ωm.

Fig. 27 Synthetic model 13

If the underground void (3 x 5 m) is placed in the conductive layer, having the same geometry, the effect is not as visible as in the previous models (Figure 28). The trend of the highest values is linear, because this result is influenced by the high resistivity of the layer not of the void.

Fig. 28 Synthetic model 14

A similar result is observed in Figures 29 and 30, where the resistivity of the bottom layer is 104 Ωm. In Figure 29, this layer is situated at 12.5m depth and in Figure 30, at 9.9m. The maximum value at the last one increases from 3 to 5 Ωm, the trend of the anomalies is similar, but the depth corresponds to the limit between the conductive and resistive layer.

Fig. 29 Synthetic model 15

Fig. 30 Synthetic model 16

If the void is placed in the resistive layer at 9.9m depth and it has low resistivity values (1 Ωm), caused by the conductive material filled, the effect is like a convex parabola, the resistive layer is significantly deformed (Figure 31).

Fig. 31 Synthetic model 17

In Figure 32 are compared the effects of two voids with similar resistivity (106 Ωm). The first is placed in a conductive medium (1 Ωm) while the second (9 x 7.5m dimensions) in a deep seated resistive layer (104 Ωm). The void placed a shallow depth (6 x 3m dimensions) gives a distinctive anomaly, while the second one effect is slightly noticeable.

Fig. 32 Synthetic model 18

In model 19 it is presented the situation of a step descendent void with the upper part in the conductive layer continuing through the resistive layer, which gives a distinctive anomaly. The void begins at a depth of 1.5m and it changes its direction slightly. The anomaly has values from 12 to 30 Ωm and corresponds to the void. There can be observed the highest values in the upper part of the anomaly, near to the surface.

Fig. 33 Synthetic model 19

The most important factors resulted to be the electrical properties of the background, the position of the underground void, its geometry and dimensions and void content (dry, water-filled/backfilled or mixture, partially or fully filled). The investigation technique settings proved to have also an influence on the resulted signal, variations of the response being observed when selecting different arrays for the same void type and configuration.

4.2.2 Results of ERT measurements in salt mines

In Figure 34 is illustrated a real example, un airfilled underground void placed at 5 m depth. The void effect is a high resistivity anomaly (6157 Ωm) , because the air is high resistive, being considered an insulator.

Fig. 34 ERT profile – Air-filled void

Another situation is an underground void filled with solid material, called also “backfilled voids”. The solid material may be homogenous or relatively homogenous. For example, an old mine entrance in mine, on ERT profile, can be observed in Figure 35.

The anomaly has minimum values (0.12-0.9 Ωm), deforms the resistive layer significantly.

Fig. 35 ERT profile – old mine entrance

The old entrance in mine can be vertical (Fig. 35) or bell-shape (Fig. 36).

Fig. 36 Bell-shape entrance in mine backfilled with inhomogeneous material

The results of an underground backfilled-void relatively homogenous are illustrated below:

Fig. 37 Underground backfilled-void relatively homogenous

The sequence of maximum-minimum values are the result of inhomogeneity of the solid material. The minimum value of resistivity (0.1-0.54 Ωm) is the effect of the contrast between the material near the surface and the old material below. The high resistivity values (15.8 – 85 Ωm), closer to the surface, are caused by the inhomogeneity of the filled material.

Chapter 5

Conclusions