Optical Radiation from an Electric Arc at Different Frequencies

The article presents research on the electric arc generated by AC current at different frequencies. The measurement procedure and system are described. Optical spectra of the generated arc in the air were recorded using a spectrophotometer. Optical spectra for five frequencies were obtained. The article also presents the energy balance of the components of the registered spectrum. The visible changes in the spectra that depend on the frequency of the AC current generating the electric arc can be significant to the diagnostics of gas insulators. The research presented in the article can be used in multiple areas of technology where an electric arc is used. The influence of the frequency of the current supplying the electric arc on the electromagnetic radiation spectrum in the area of light radiation emitted by the electric arc allows for the construction of systems that can shape the desired characteristics of the electric arc.


Introduction
The electric arc generated in gas insulators is the penultimate phase of the electrical discharge and leads to a breakdown. The order in which this happens is corona extinction voltage (CEV); partial discharge initiative voltage (PDIV); partial discharge extinction voltage (PDEV); arcing voltage; breakdown voltage. There are many mechanisms to stop the formation of an arc [1,2] and prevent a breakdown. Just as important as the breakdown is the diagnostics of the gas isolation state in the direction of partial discharges (PDs) and corona discharges, which are the phenomena preceding the formation of an arc. Many diagnostic methods are used for this purpose: acoustic emission (AE), electromagnetic wave detection in the UHF range, high energy ionizing and UV or visible light (with a spectrophotometer). The main aim of the research presented in this article was to determine and analyze the spectrum of optical radiation in the ultraviolet, visible and near-infrared range (UV-NIS-NIR) emitted by a generated electric arc in air at atmospheric pressure for different AC voltage frequencies.
The recording of the arc spectrum is one of the elements of high-voltage diagnostics. It is possible to define some descriptors for a certain phenomenon by their shape or intensity. From the scientific point of view, for different frequencies of alternating voltage generating an electric arc, the spectrum of the arc differs depending on the preset frequency. Characterization of parameters for different frequencies of currents may help us to identify risks depending on the frequency of current used. Previous research on electric arcs was more focused on phenomena related to direct current. Guan et al., in the article "DC arc self-extinction and dynamic arc model in open-space condition using a Yacob Ladder" [3], focused on the study of the arc generated at the initial current from 50 A to 200 A, while the voltage of the generating system was 560 V from a three-phase system. The entire process of arc evolution, from its production to development to self-extinguishing, was studied. The phenomenon was recorded with a fast camera in the visible light range. A Yacob Ladder was used in the study. Another study, in which a Yacob Ladder was used to generate the arc, is called "Application of Optical Spectrophotometry for Analysis of Radiation Spectrum Emitted by Electric Arc in the Air" [4]. The authors also used DC current. However, the arc measurements were performed with an optical method, using a spectrophotometer in the visible and near UV light range. Other studies aimed at characterization of the electric arc are presented in a paper by Martins et al. [5]. The authors used a high-speed camera to record the optical signal. In this work, different levels of current peaks are used, from 10 kA to 100 kA, with a short peak duration of about 15 µs. The camera itself is synchronized with the trigger of the arc generator. The ionic lines of nitrogen and oxygen are used to determine the radial temperature profiles and electron density in the arc channel over a period spanning from 2 µs to 36 µs. Tests in contaminated isolators using the optical method were described in a paper titled "Study of the AC arc discharge characteristics over polluted insulation surface using optical emission spectroscopy" [6]. In other tests, the arc propagation growth and its shape and leakage current at various air pressures were checked. The length of the discharge path is related to air pressure and it is always shorter with decreasing pressure [7]. A description of arc formation and propagation is also included in the study "Performance and Characteristics of a Small-Current DC Arc in a Short Air Gap" [8]. This paper shows that the color of the DC arc changes from blue to purple and then yellow, forming a flame in the visible light range as the current increases. Research on the electric arc with diagnostic methods for PD detection has been described in an article by Chen et al. [9]. The authors simulated the wave of acoustic pressure and then carried out experiments demonstrating the convergence of theory and practice. It was found that the spatial distribution of power density in the arc is highly heterogeneous and the power density in the area close to the electrodes is much higher. The acoustic method is widely used for PD detection in both gaseous and electrical insulating liquids [10,11]. The phenomenon of acoustic wave formation itself, at the moment of arc generation, can be very dangerous. High pressure and high energy can cause a lot of damage and can be dangerous for life [12]. As mentioned in the article by Martins et al. [5], the electric arc excites atoms of ambient elements. The optical spectrum recorded in this study was based on the nitrogen and the oxygen spectra. With direct current, the excitations change only depending on the energy supplied to generate the arc. Extended studies, with currents from 100 kA to 250 kA at the peak, were presented in [13]. In the case of sine wave currents, the excitations can look different. For different frequencies, different elements may be induced into resonances at the nuclear level at different instances of the phenomenon. The excitation energy concerns mainly electrons, but it may also affect the nuclear spin of the elements. Non-zero nuclear spin includes almost all atoms with an odd number of nucleons (e.g., hydrogen 1 H, carbon 13 C, nitrogen 15 N, oxygen 17 O, fluorine 19 F, sodium 23 Na and phosphorus 31 P). Put simply, a nuclear spin can be imagined as the rotation of the nucleus around its axis. It is related to the internal momentum of the nucleus. The main purpose of the research presented in this article was to determine the spectrum of light in the visible range coming from a generated electric arc in air at atmospheric pressure, depending on the frequency of alternating voltage. In the study, an optical method was used to detect and identify partial discharges in insulators of equipment and power cables [14]. This method is often combined with others using different ranges for the detection of electromagnetic radiation emitted from PDs [15,16]. Research on arcs and preceding corona discharges (CDs) is also important in aeronautics. The electric arc is a major risk to aircraft systems. Professor Riba's team was engaged in diagnostics of corona discharges under aeronautical conditions. In the article [17], they described an experiment using a low-cost camera for early detection of UV radiation from corona discharges. They also compared the corona discharge measurements for positive and negative DC and AC for 50 Hz. Experimental results presented in [18] clearly show that the sphere-plane gap follows a correlation similar to Peek's law for cylindrical conductors. This conclusion is true for 50 Hz AC, positive DC and negative DC supply. However, for different AC frequencies, the UV signals from CDs may be different. This is important for producers of HVAC equipment. Jiang and others used an optical method to track DC discharge signals [19]. They found that this research is important for systems of more electric aircraft (MEA). Experimental results indicate that the MEA DC system has several series of characteristic arc spectra, including 309.3 and 324.5 nm in the UV range. The anode material is also relevant for the spectrum range. The advantage is immunity to electromagnetic interference and a quick reaction to changes in the studied phenomenon. However, Jiang's article lacks information about research on AC arcs, especially for different current frequencies. Electromagnetic wave detection in the field of visible and UV light has been used since the 1980s. The article [20] proposed research on the spectrum of light coming from aluminum and copper as the main materials that can become electrodes during arc generation. Significant activity of ionized nitrogen atoms was shown, which indicated that they could be potential spectrum signals for arc flash detection operations. An Ocean Optics optical spectrophotometer was also used in this study.
Other diagnostic methods for the detection of electrical discharges used, for example, in transformers or overhead power lines may also be helpful. Under aeronautic conditions, however, the most accurate methods are be those based on EM radiation detection. Special attention should be paid to the detection of high-energy radiation [21] and the very popular UHF method [22,23], which extend the range of EM wavelength detection to longer and shorter wavelengths.

Measurement System and Methodology
The measurement system ( Figure 1) consisted of an arbitrary generator (Tektronix AFG1022, Tektronix, Beaverton, OR, USA), whose signal output was connected to the input of a power amplifier based on the Texas Instruments OPA541 operational power amplifier. The power amplifier's signal output was connected to the primary winding terminals of a high-voltage transformer with a ferromagnetic core of K = 1/500 ratio, whose secondary winding terminals were connected to a spark gap. The spark gap was placed together with the light spectrum analyzer probe in a non-transparent body. The optical output of the light spectrum analyzer probe was connected to the first pin of the polymer optical fiber (POF), whose second pin was connected to the light spectrum analyzer-an HR4000 high-resolution optical signal processing spectrophotometer. The communication port of the light spectrum analyzer was connected to the communication port of a PC. including 309.3 and 324.5 nm in the UV range. The anode material is also relevant for the spectrum range. The advantage is immunity to electromagnetic interference and a quick reaction to changes in the studied phenomenon. However, Jiang's article lacks information about research on AC arcs, especially for different current frequencies. Electromagnetic wave detection in the field of visible and UV light has been used since the 1980s. The article [20] proposed research on the spectrum of light coming from aluminum and copper as the main materials that can become electrodes during arc generation. Significant activity of ionized nitrogen atoms was shown, which indicated that they could be potential spectrum signals for arc flash detection operations. An Ocean Optics optical spectrophotometer was also used in this study. Other diagnostic methods for the detection of electrical discharges used, for example, in transformers or overhead power lines may also be helpful. Under aeronautic conditions, however, the most accurate methods are be those based on EM radiation detection. Special attention should be paid to the detection of high-energy radiation [21] and the very popular UHF method [22,23], which extend the range of EM wavelength detection to longer and shorter wavelengths.

Measurement System and Methodology
The measurement system ( Figure 1) consisted of an arbitrary generator (Tektronix AFG1022), whose signal output was connected to the input of a power amplifier based on the Texas Instruments OPA541 operational power amplifier. The power amplifier's signal output was connected to the primary winding terminals of a high-voltage transformer with a ferromagnetic core of K=1/500 ratio, whose secondary winding terminals were connected to a spark gap. The spark gap was placed together with the light spectrum analyzer probe in a non-transparent body. The optical output of the light spectrum analyzer probe was connected to the first pin of the polymer optical fiber (POF), whose second pin was connected to the light spectrum analyzer-an HR4000 high-resolution optical signal processing spectrophotometer. The communication port of the light spectrum analyzer was connected to the communication port of a PC.
The HR4000 high-resolution optical signal processing spectrophotometer, manufactured by OceanView, with a spectral range from 200 nm to 1100 nm, was used to record the emitted optical radiation.  The HR4000 high-resolution optical signal processing spectrophotometer, manufactured by Ocean Optics (Edinburgh, UK), with a spectral range from 200 nm to 1100 nm, was used to record the emitted optical radiation. In a powered system, an electric arc was initiated in the spark gap. Then, in the settings panel of the arbitrary generator, a preset frequency of sinusoidal waveform generation was set, with a fixed amplitude of the generated high voltage supplying the spark gap (about 5 kV). In the next step, the spectrum emitted by the electric arc of light was measured and the results, along with the preset frequency of the waveform supplying the spark gap, were recorded by a PC.
The measurements were carried out in a laboratory under constant metrological conditions. The measuring system was placed in a darkened room without external light sources that could disturb the measurements. In addition, a calibration of the background compensation was performed before each measurement test. The electric arc was generated at alternating voltage U = 5 kV for the following frequencies: 13.5 kHz, 20.0 kHz, 80.0 kHz, 100.0 kHz and 150.0 kHz. The air temperature was constant at 20 • C and the humidity was also constant at 48%. Due to the experimental nature of the study, the influence of temperature and humidity on recorded values was not analyzed at this stage.
In the quantum description, the components of the wavelengths are marked as a photon stream, where each wavelength of emitted radiation corresponds to an energy quantum, i.e., a photon of a specific energy. The energy of this photon (E) can be determined from the following equation: where E is the quantum energy (J), h is Planck's constant (6626·10 −34 (J·s)) and υ is the wave frequency (1/s). Wave frequency is expressed in the relation: where: υ is the wave frequency (1/s); c-phase speed-is the speed of light in a vacuum (2998·108 (m/s)); and λ is the wavelength (nm). The analysis of the obtained spectral distributions showed that, on the basis of Equation (1), it is possible to determine the energy of optical radiation by taking into account the number of photons for particular wavelengths. The number of counts was determined on the basis of the intensity of individual wavelength components recorded by the spectrophotometer, where a single count corresponds to a certain number of emitted photons. The number of photons per count was determined on the basis of the technical parameters of the used optical spectrophotometer. Finally, Equation (1) can be expressed as follows: where E is the quanta energy (J), n is the number of photons per one count (-), h is Planck's constant (6626-10-34 (J-s)) and υ is the wave frequency (1/s). Figure 2 presents the results obtained from the measurements in the form of optical spectra, which were recorded for individual frequencies of electric-arc-generation voltage. The presented spectra were determined on the basis of averaged measurements of wavelength components from the applied UV-NIS-NIR optical radiation range. The spectra were formed as a result of photon emission caused by the electric field.

Results and Discussion
The analysis of the results indicates the repeatable nature of the recorded optical spectra for specific supply voltage frequencies. Spectral characteristics are composed of band and continuous spectra and their dominant intensity is in the ultraviolet range ( Figure 3). The dominant components of wavelengths and spectral ranges were determined on the basis of the obtained spectral characteristics. They are presented in Table 1.     Registered emission spectra in the UV-NIS-NIR range can be qualified as continuous-band spectra due to their shape. The continuous spectrum is characterized by the occurrence of sequentially ordered wavelength components in a continuous manner, which is a characteristic feature of liquid and solid emissions. In contrast, the band spectrum is created by combining several individual components of the wavelength, which indicate the activation of different elements. Table 2 shows the optical energy balance for each range and estimates the total energy for the analyzed UV-NIS-NIR range. Results are presented as average values from an individual series of 50 partial measurements.  For the particular frequencies of the arc supply voltage, the percentage share of energy for particular ranges of optical radiation was determined and is shown in Figure 4. Registered emission spectra in the UV-NIS-NIR range can be qualified as continuous-band spectra due to their shape. The continuous spectrum is characterized by the occurrence of sequentially ordered wavelength components in a continuous manner, which is a characteristic feature of liquid and solid emissions. In contrast, the band spectrum is created by combining several individual components of the wavelength, which indicate the activation of different elements. Table 2 shows the optical energy balance for each range and estimates the total energy for the analyzed UV-NIS-NIR range. Results are presented as average values from an individual series of 50 partial measurements. For the particular frequencies of the arc supply voltage, the percentage share of energy for particular ranges of optical radiation was determined and is shown in Figure 4.
The obtained energy values in this case were only intended to be used to analyze the contribution of individual ranges of optical radiation. Determined percentage distributions showed the predominant proportion of ultraviolet radiation for all analyzed voltage frequencies of the generated arc. It was also observed that, with the increase of the voltage frequency at which the arc is generated, the near-infrared radiation fades. This situation may be caused by the flow of higher current through the electric arc at lower frequencies that, in turn, increases the temperature of electrodes. Metals are not luminescent materials; under the influence of a significant local temperature increase, which approaches their melting point, they emit a near-infrared component [24]. However, the emission of radiation in the ultraviolet range results from the glow-type character of the electric arc and is caused by the recombination of the released electrons through electroluminescence [25]. The obtained energy values in this case were only intended to be used to analyze the contribution of individual ranges of optical radiation. Determined percentage distributions showed the predominant proportion of ultraviolet radiation for all analyzed voltage frequencies of the generated arc. It was also observed that, with the increase of the voltage frequency at which the arc is generated, the near-infrared radiation fades.
This situation may be caused by the flow of higher current through the electric arc at lower frequencies that, in turn, increases the temperature of electrodes. Metals are not luminescent

Conclusions
The main aim of the research presented in this article was to determine the spectrum of light in the visible range coming from a generated electric arc in air at atmospheric pressure, depending on the frequency of alternating voltage. It was noted that, with the increase in the frequency of the current generating the arc, the percentage of the UV component increased at the expense of other components. Increasing the percentage of UV radiation in air ionization results in the formation of more ionized atoms and faster changes in the composition of gas insulators; it also reduces its electrical strength, resulting in easier formation of plasma channels. This is important in the study of partial discharge spectra in SF6-based insulators. The spectrum of a mixture changes depending on its structure. Studies of the decomposition of these insulators were carried out by Dincer et al. [26]. The next steps for the authors of this paper are to study the spectra in other gas insulators and to associate them with the composition of the mixture at different stages of the latter's aging.
The research presented in the article can be used in multiple areas of technology where electric arcs are used. The influence of the frequency of the current supplying the electric arc on the electromagnetic radiation spectrum in the area of light radiation emitted by the electric arc allows for the construction of systems that can shape the desired characteristics of the electric arc. For example, a simple system with a selective light detector (or set of detectors) with a specific wavelength in a feedback system that affects the frequency of the inverter of the high-voltage power supply feeding the arc can be used to upgrade and optimize the operation of existing arcing equipment. Examples of applications of such systems, which can translate into measurable benefits, such as increasing the efficiency of equipment and processes or reducing electricity consumption, are: metallurgical furnaces and welding equipment, where optimization consists in increasing the light emission in the infrared area; high-efficiency arc lamps, where it is profitable to minimize light components in the infrared area and intensify them in the visible and ultraviolet areas; and in chemical synthesis devices, where the area of optimization of the emission of selected spectrum components should be selected individually for specific chemical processes.

Conflicts of Interest:
The authors declare no conflict of interest.