**3. Results**

The results of the experiment are discussed in two subsections. The first part presents the magnetic field distribution measurements using MO-sensor and their application for rebars detection. The experiments can be assumed as preliminary studies. The measurements with the MO-sensor are carried out quickly, and they are easy to interpret. However, the sensitivity of the MO-sensors is lower than the AMR sensor, and there is no possibility to measure *x*, *y*, and *z* induction components. In this case, only sample S1 with single rebar is tested (all samples were tested with the AMR sensor as shown in the following subsection). The experiments with the MO sensor show both the influence of magnetization on increasing the ability to detect rebars and the application potential of the MO-sensors. The same pole magnetization (SPM) is used in this experiment.

In the following subsection, results received for three different samples and three different magnetization variants are presented. All experiments were conducted with the same magnets (having different orientations against the rebars). Therefore, the magnetization effect is weaker for samples with a bigger concrete cover thickness. In addition, always the same single AMR sensor was used. The main point of the experiments is to show the impact of the magnetization method on received results. Tests prove that the impact is even more significant for more complex samples. Experiments carried out on the samples simulating reinforced mesh (samples S2 and S3) showed that the CMT (Continuous magnetization techniques) were much more effective than RMT (residual magnetization techniques). Moreover, the SPM (same pole magnetization) allows identifying rebars more straightforwardly than OPM (opposite pole magnetization).

### *3.1. Experiments with the MO-Transducer*

Experiments using the MO-sensor for sample S1 (with single rebar) were conducted to show the differences between CMT and RMT. Rebars are magnetized every time up to the same level and in the same orientation. The SPM was selected as a method of magnetization. As a reference, the same experiment was also conducted with the non-magnetized rebar. Experiments were taken with the step of 5 mm along the axis *z* (change of concrete cover thickness *h*), and 20.5 mm along the axis *x* (size of the sensor is 15.5 × 20.5). In this way, continuous measurements were obtained without any gaps. The thickness of the concrete cover *h* was changed in the range from 0.5 to 100 mm.

Predictably, experiments have shown that magnetized rebar can be detected with a much greater concrete cover than a non-magnetized. When the non-magnetized rebar is

challenging to detect with a cover thickness above *h* = 20 mm, the magnetized rebar could be detected from a distance of more than 100 mm. However, the readability of the graphs for large cover thicknesses is limited. Examples of the measurements received for thick concrete cover are shown in Figure 13. Only half of the measurement results are shown (the other half is symmetrical).


**Figure 13.** Magnetic field distribution measured with the MO-sensor for different concrete cover thicknesses; same pole magnetization; only half of the measurements are shown.

Plots showing the magnetic field distribution over the magnetized rebar vary depending on the thickness *h* of the concrete cover. Examples of such characteristics are shown in Figure 14. They are repeatable and unambiguous. Therefore, on their basis, it is possible to estimate the location of the rebar, the thickness of the concrete cover, and possibly other parameters of the structure, as was the case in [9,10].

In existing civil engineering constructions, the thickness of the concrete cover is usually between 10 mm to 50 mm over the reinforcing bars. When the reinforcing bars are not magnetized, the MO sensors are not sensitive enough to detect rebars from such distances. However, when the bars are magnetized, the efficiency of the MO sensors is sufficient. Thus, sensors of this type are suitable for the CMT and not for RMT. Examples of calculated signal to noise ratio (SNR) values are presented in Table 1.

**Table 1.** Signal to noise ratio SNR calculated for measurements obtained by MO sensor for different concrete cover thicknesses *h*.


**Figure 14.** Impact of concrete cover thickness on the MO-sensor measurements. The average line profile of the magnetic field was measured using MO-sensor with the same pole magnetization SPM by moving the sensor along the *y*-axis.

The SNR was defined as:

$$\text{SNR} = 20 \log \frac{A\_{\text{sigmal}}}{A\_{\text{noise}}}$$

#### *3.2. Influence of Rebars Magnetization Method on Magnetic Field Distribution*

All experiments in this section were taken with the step of 5 mm along the axis *z* (change of concrete cover thickness *h*), from 20 to 70 mm (typical concrete cover thickness).

The step along the axis *y* was equal to 20 mm and experiments were taken from −100 to 100 mm. Position 0 is a position in the middle of the rebar.

The step along the axis *x* was equal to 2 mm and experiments were taken from 0 to 98 mm. Rebars in S1 and S2 are placed in position 27 mm (axis *x*). In the case of S3, the middle rebar is placed in this position.

Magnets were moving together with the sensor and were placed on both sides of the sensor at a distance of 500 mm.

In the first set of experiments, the measurements were carried out for the sample S1 using different magnetization methods. The results for different thicknesses *h* of concrete cover are presented in Figure 15. The second set of experiments was carried out with three different samples S1, S2 and S3, shown in Figure 3. The measurement results were symmetrical concerning the rebar, and therefore the measuring range has been reduced nearly by half. Positions of the rebars were depicted on the plots in Figure 16 by dashed lines.

Figure 15 shows that the influence of magnets on the rebar decreases when the cover thickness *h* is increasing, and thus, the magnetic field measured by the sensor also decreases. The method of magnetization significantly influences the value of the magnetic field. Compared to the field measured for a non-magnetized bar, the use of magnetic excitation in any configuration of the magnets causes an increase in the field value. As a result, the magnetic field diagrams obtained for different cover thicknesses *h* differ significantly, which facilitates identification. The strongest field over the rebars was measured in the case of magnets directed towards the bar with homonymous poles (SPM), lower for magnets with opposite poles (OPM), and the lowest for the reference sample in which the rebar was not magnetized. One can observe that the maximum value of the magnetic field component *Bz* was similar in both magnetization methods.

In the case of non-magnetized rebar, the graphs representing the magnetic field along the *x*-axis perpendicular to the rebar did not change significantly with increasing cover thickness. Even the changes measured for thickness *h* above 50 mm are minimal.

**Figure 15.** *Cont*.

**Figure 15.** Results of 2D measurements using the AMR-sensor obtained for different variants of magnetization and different thicknesses *h* of concrete cover; experiment conducted for the sample S1 with single rebar; SPM—same pole magnetization, OPM—opposite pole magnetization; concrete cover thickness: (**a**) *h* = 20 mm; (**b**) *h* = 30 mm; (**c**) *h* = 50 mm; (**d**) *h* = 70 mm.

**Figure 16.** Selected results of 2D measurements of magnetic field component *Bx* with the depicted position of the rebar (dashed line); (**a**) sample S1—single rebar; (**b**) sample S2—two rebars one under the other; (**c**) sample S3—three rebars next to each other.

The results obtained with the SPM magnetization system are the easiest to interpret. The *Bx* component of the magnetic field is particularly interesting. It has a much larger value than the others and changes significantly with increasing cover thickness. Moreover, in contrast to magnetization OPM, the SPM looks similar, regardless of the measurement place in the *y* axis direction. The most important conclusion from the presented results is that the cover thickness *h* can be estimated based on the slope of the graph of the measured magnetic field (Figure 17).

The measurements show that the lack of magnetization causes a significant reduction of the magnetic field and, therefore, it may cause errors in the rebars identification. For example, in Figure 17, in the case of non-magnetized rebar, the *By* component takes very small values. The results of experiments show that the magnetization method can impact noise immunity. Signal to noise ratio calculated for different methods of magnetization and different thicknesses of concrete cover *h* is provided in Table 2.


**Table 2.** Signal noise ratio SNR (dB) calculated for measurements obtained using AMR sensor for different concrete cover thicknesses *h*, and different magnetization methods.

**Figure 17.** Plots of the magnetic field components as a function of sensor position *x* (mm) which are showing the influence of the concrete cover thickness *h*.

As mentioned, the signal to noise ratio (SNR) for *y*-component and non-magnetized rebar has very small values. Moreover, the presence of rebar nearby does not appear to be the dominant factor that formed this characteristic (due to the influence of external fields). Therefore, SNR is not calculated in that case. The impact of noise is much higher in the case of non-magnetized rebars. A slightly bigger SNR was achieved for the SPM magnetization compared to the OPM. However, in this respect, both methods are comparable. As it is not difficult to predict, the growth of the thickness of concrete cover has a negative effect on the SNR. The influence of *h* on the SNR is different for different magnetization methods and for different components. However, drawing conclusions based on Table 2 could be premature due to a small test attempt. In addition, in all cases, the impact of noise is moderate. It can be noted that the dominant influence on SNR has the maximum value of the obtained signal. The relationship between the signal value received from the AMR sensor and the thickness of the concrete cover *h* for different magnetization methods is shown in Figure 18.

**Figure 18.** Graphs of the maximum value of the magnetic field components as a function of the concrete cover thickness *h* obtained for various magnetization methods.

The maximal value of signals presented in Figure 18 is greater in the case of the *Bx* and *Bz* for SPM (same pole magnetization). In the case of *By*, the biggest signal value is achieved for OPM (opposite pole magnetization). One can observe that the maximal signals are significantly smaller without magnetic excitation. Results of identification, in that case, are uncertain (nevertheless, detection of the reinforcement is possible). The problem of measurements conducted without magnetization is the low value of the received signals. This problem causes strong noise influence and the characteristics ambiguity. It is worth noting that the maximum amplitude of various components is significantly different. The maximal value of the signal obtained for *Bz* is much higher than for the two other components. This fact has a substantial impact on the SNR. The comparison of the characteristics for the non-magnetized rebars, magnetized with the SPM and magnetized with the OPM, is shown in Figure 19 (normalized curves). For non-magnetized rebars, there are differences in the shape of the characteristics caused by noise and the ambiguous polarity of the rebars. As a result, they are challenging to interpret, and the identification results could be inaccurate. Differences between maximum values of signals obtained for SPM and OPM are minor. Obtained results in these two cases are comparable.

**Figure 19.** Graphs of the normalized maximum values of the magnetic field components as a function of the concrete cover thickness *h* obtained for various magnetization methods.

The type of magnetization method does not affect the steepness of changes in the measured linear profiles of the magnetic field components (Figure 20). However, also in this aspect, low SNR makes identification difficult in the case of lack of magnetization. The curves obtained for SPM and OPM are almost the same.

The following experiment was carried out to investigate the influence of magnetization on identifying the reinforcement mesh. Two kinds of specimens were tested: sample S2— (two rebars one over the other—Figure 15b) and sample S3 (three rebars are next to each other—Figure 15c) are considered in the tests and compared with measured earlier sample S1. The results are presented in Figure 21.

**Figure 20.** Graphs of the normalized value of the magnetic field component *Bx* (*x*) as a function of the concrete cover thickness *h* (mm) obtained for various magnetization methods.

**Figure 21.** Results of 2D measurements using AMR sensor obtained for different samples (S1, S2, S3); without magnetization; thickness of the concrete cover *h* = 40 mm; *x* (mm), *y* (mm)—sensor positions.

The obtained results indicate that regardless of the method of magnetization or the lack of it, more complex structures containing several bars next to each other (samples S2 and S3) generate field distributions significantly different than in the case of a single bar (sample S1). Correctly-configured magnetic excitation creates opportunities to correctly identify complex structures, which are more similar to existing building structures.

There are many problems with testing reinforcement meshes, where more than one rebar strongly influences the sensor. In the case of concrete structures without magnetization (RMT) the most significant problem is a lack of knowledge about the residual magnetization of individual rebars. Another obstacle is that the rebars could be strongly magnetized during earlier operations (e.g., by a crane with an electromagnetic gripper) and

the obtained results strongly depend on the magnetized rebars relative position as shown in Figure 22.

**Figure 22.** (**a**) Polarizations (residual magnetization) of earlier magnetized rebars in the sample S2; (**b**) Results of 2D measurements using AMR sensor received for different arrangements of earlier magnetized rebars; without external magnetization during measurements, the thickness of the concrete cover *h* = 40 mm; sample S2; *x* (mm), *y* (mm)—sensor positions.

The problem with unknown residual magnetization disappears when magnetic excitation is used. Moreover, the signal value is higher and the identification process is reliable. Next, experiments were conducted for three different samples with the use of different magnetization methods.

In the case of SPM, the value of the obtained signal is bigger than without the magnetization. Unfortunately, identifying the arrangemen<sup>t</sup> of the bars in the mesh is very difficult or even impossible. The shapes and maximal values of received characteristics are very similar for sample S2 and sample S3, as shown in Figure 23.

**Figure 23.** Results of 2D measurements using AMR sensor received for different samples (S1, S2, S3); the magnetization OPM; the thickness of the concrete cover *h* = 40 mm; *x* (mm), *y* (mm)—sensor positions.

Experiments prove that SPM is far superior to OPM in identifying complex structures. For sample S2, in which the bars are located one after the other, the characteristics are to some extent similar to those obtained for single rebar. However, their shapes and maximal values differ enough, and they are easy to distinguish. Therefore, it is possible to easily recognize this arrangemen<sup>t</sup> of rebars and even estimate the distance between them. In the case of sample S3, where the rebars are next to each other, the greatest signal values are obtained over the middle rebar (over which the magnets are placed). In addition, this case is easy to recognize. The SPM-results are presented in Figure 24.

**Figure 24.** Results of 2D measurements using AMR sensor received for different samples (S1, S2, S3); the magnetization SPM; the thickness of the concrete cover *h* = 40 mm; *x* (mm), *y* (mm)—sensor positions.

### **4. Discussion**

The use of magnetic excitation is crucial for the quality of the results in the magnetic evaluation of reinforced concrete structures. In the case of simple structures, where only one rebar is detectable, it affects noise immunity (Table 2) and the signal value. In addition, even a weak magnetic field makes the rebar's polarization predictable, which significantly facilitates identification. As shown in Figure 19, the results of measurements obtained without magnetization are challenging to predict and heavily dependent on residual magnetization (which can be unknown to the investigator). Generally, the identification of any parameters without magnetic excitation is a subject of significant uncertainty. However, it is possible to detect the rebar even without the magnetization. In the case of more complex structures (Sample S2 and S3), identifying the structure can be tricky when two or more rebars of unknown polarization affect the sensor.

The thickness of the concrete cover (*h*) can be estimated using the magnetic method. The relationship between the signal value and the *h* for different magnetization methods is shown in Figure 18. Potentially also different parameters of a reinforced concrete structure can be tested with this method (e.g., rebars diameter, rebars class, etc.). However, confirmation requires further investigations.

The magnetization method significantly impacts the results of measurements performed with the magnetic method. This aspect is often undervalued. In the case of sample S1, signal value and SNR depend on magnetization methods. Better results are received mostly for SPM (single pole magnetization). Moreover, in the case of the SPM, identification was more straightforward, as the results received for *B*x are similar over the entire surface above the rebar (Figure 15). The magnetization method is even more critical in evaluating

more complex structures. In the case of samples S2 and S3, it was possible to identify the structure only by using the SPM (Figures 23 and 24).

The MO sensors enable the evaluation of large areas of reinforced concrete structures in real-time. It is also helpful for fast pilot studies. In the case of greater concrete cover thicknesses, it is necessary to magnetize the rebars due to the moderate sensitivity of the MO-sensor. The signal to noise ratio (SNR) in the case of MO-sensors is much lower than in the case of the AMR sensor. Therefore, for more accurate tests, MO-sensors are not well suited. However, the quality of the results can be improved by hardware enhancement.

The AMR sensors enable effective testing of reinforced concrete structures without magnetization (with typical concrete cover thickness). However, when the concrete cover thickness is high, it is worth using even a small level excitation to improve the system's efficiency. This solution provides a stronger signal, easier to interpret and analyze the characteristics.

MR elements can be used for area testing. For this purpose, matrices of the sensors can be used. The experiments presented in the paper show that the elements of this kind are much more sensitive and resistant to noise than MO. Comparison is presented in Table 3. The MR sensors also allow the testing of particular spatial components *Bx*, *By*, *Bz*. However, these elements cannot be used if the magnetic field is out of range. Therefore, if the magnetic field can be stronger than 1 mT, it is recommended to use a proper MO-sensor or matrix of Hall-elements (Hall-elements possess all advantages of MR sensors, but the active range is much higher than in the case of AMR and the sensitivity is comparable with MO-sensors).


**Table 3.** Signal to noise ratio SNR calculated for measurements obtained by MO and AMR sensors for different concrete cover thicknesses *h*, and same pole magnetization.
