Series Arc Fault Characteristics and Detection Method of a Photovoltaic System

Abstract

The DC arc is the main cause of fire in photovoltaic (PV) systems. This is due to the fact that the DC arc has no zero-crossing point and is prone to stable combustion. Failure to detect it in a timely manner can seriously endanger the PV system. This study analyzes the influences of the series arc and the maximum power point tracking (MPPT) algorithm on the PV output characteristics based on the PV equivalent circuit module. The PV voltage and current variation characteristics are obtained when the series arc occurs. The findings indicate that the input voltage of the converter remains unchanged due to the MPPT algorithm before and after the series arc occurs. Furthermore, the PV faulty string output current will drastically decrease when the series arc fault occurs. On this basis, a series arc detection method based on the current change is proposed, suppressing the combustion of the series arc by increasing the target voltage of the MPPT algorithm. The experimental results show that the proposed method can effectively detect and extinguish the series arc in the PV system within 0.6 s. Compared to the other methods, the proposed method can be integrated into the PV system without additional hardware.

1. Introduction

With the increasing severity of environmental pollution, PV power generation systems have been developed rapidly because they are clean and renewable, resulting in an increasing number of unexpected failures in PV plants. Loose inner electrical cables or degradation of the cable insulation may cause a series arc fault to occur [1]. DC arc faults always involve high temperatures and do not extinguish themselves because of the arc current’s absence of a zero-crossing point [2], which can easily cause stable combustion, leading to fire hazards. Therefore, the National Electrical Code stipulates that arc identification and protection devices for all PV systems have a voltage of over 80 V [3].

Arc faults can be divided into three categories: parallel arc, series arc, and ground arc [4]. The ground arc and parallel arc can be easily identified through general protections because of strong changes in the loop current [5]. The conventional protections make it hard to identify the series arc because the bus voltage and loop current will not increase in the PV system [6]. Therefore, relevant research mainly focuses on identifying the series arc fault. These researchers typically use three methods to analyze fault signals: time-domain methods, frequency-domain methods, and hybrid methods.

The time-domain method mostly detects the arc fault utilizing the fault characteristics in the time domain. In [7], the PV panel current entropy was utilized to recognize the series arc. In [8], the supply voltage and line current were used to identify the series arc. In [9], the signal-to-noise ratio was utilized to identify the PV series arc and remove switching signatures based on cross-correlation. In [10], the covariance matrix was used to separate the voltage and the current, and the fault characteristic index was obtained using the principal component analysis. In [11], the signal’s degree of predictability or certainty was used to identify the PV series arc. In [12], the mathematical morphology was utilized to recognize the PV series arc. However, the change in light intensity and the disturbances of their controllers will also change the time-domain current characteristics and lead to the maloperation of these methods. In [13], the voltage drop was utilized to recognize the series arc and extinguish the series arc by minimizing the duty cycle of the converter. However, it did not consider the effect of the MPPT method, and minimizing the duty cycle may cause system instability.

The frequency-domain method utilizes the fault characteristic in the frequency domain to recognize the arc fault. A common method for analyzing the signal in the frequency domain is fast Fourier transform (FFT). In [14], the obvious discrepancy in the frequency domain between the arc current and the normal current was utilized to recognize the arc. However, if the considered bandwidth has switching noises, it may lead to the maloperation of the detection method based on a single frequency band. To address this problem, an identification method based on multiple frequency bands was proposed [15]. In [16], a series arc detector using the FFT analysis method was verified in the photovoltaic system. In [17], the relative magnitude comparison was utilized to recognize the series arc. In [18,19], the current of parallel capacitors was utilized to recognize the PV arc fault, but the required economic costs increased. In [20], a magnetic sensor was used to recognize the series arc by measuring the pink-noise signal. In [21], the high-frequency spectral pattern was utilized to identify the PV series arc using ideas from compressed sensing. In [22], the low-frequency spectrum was utilized to identify the series arc. However, FFT analysis cannot obtain the time-domain characteristics and it is sensitive to switching noises, which may lead to the maloperation of these methods.

The hybrid methods mostly detect the series arc fault using signal characteristics in the time and frequency domains. In [23], the erratic fluctuation and high-frequency noises were chosen as the series fault characteristics to diagnose the series arc. A series arc diagnostic method uses the discrete wavelet transform (DWT) method to acquire the fault characteristic of the current in the time–frequency domain [24]. In [25], an arc fault diagnostic method integrating DWT, zero-range density analysis, and periodic feature analysis was proposed. In [26], the voltage filtered via the Kalman filter was used to recognize the PV series arc. In [27], the adjacent multi-segment spectral similarity characteristics of the signal were utilized to recognize the series arc. The variational mode decomposition method was used to extract the arc characteristics [28]. In [29], an arc model was established based on the V-I, power spectra, and current drop characteristics of the series arc, and an arc diagnostic method based on autocorrelation and magnetic field sensing was proposed. However, these methods require the manual selection of thresholds, which determine the accuracy of these detection methods. Recently, some detection methods based on machine learning have been studied. In [30], a series arc diagnostic method integrating multi-scale fuzzy entropy, adaptive local mean decomposition, and a support vector machine (SVM) was proposed. The optimized variational mode decomposition and the SVM were utilized to recognize the PV series arc [31]. In [32], the lightweight convolutional neural network was utilized to diagnose the PV series arc. In [33], a transfer-learning-based detection network was utilized to identify the series arc through low energy. In [34], a series arc diagnostic algorithm based on ensemble machine learning was proposed. However, these methods need much experimental data to train the machine-learning model. These models are trained using data obtained within a specified system, but these trained models have low accuracy for untrained systems.

At present, research on the series arc detection method mainly utilizes the high-frequency noise of the voltage and current signal. Therefore, these methods require additional hardware and complex algorithms to detect the series arc, increasing the cost of PV systems.

This study analyzes the influence of the series arc and the MPPT algorithm on the PV output characteristics. On this basis, a series arc detection and extinguishing method is proposed. The method detects the series arc using the current drop and extinguishes the arc by increasing the bus voltage. Then the results indicate that the proposed method can successfully detect and extinguish the series arc in the low-voltage PV system without the additional hardware and can be verified in a PV system with the UL1699B standard [35].

2. The Influence of a Series Arc on the PV System

The typical structure of the two-stage inverter in the PV power generation system is shown in Figure 1 [36]. The first stage is a DC–DC converter, which achieves voltage boosting and MPPT function. The second stage is a DC–AC converter, which achieves the inverter and grid connection function, where I_loop and U_bus are the loop current and bus voltage. U_dc and $U_{\mathrm{d}\mathrm{c}}^{*}$ are the DC bus voltage and its target value. i_o and u_o are the current and voltage of the grid side.

The series arc may occur at various positions, as depicted in Figure 1: on the main PV bus (#1), on the PV string (#2), and between the PV module (#3).

The series arc that occurs on the PV array has less impact on the DC–AC converter and its control. Therefore, this study mainly analyzes the impact of a series arc on a PV array and boost converter with its control method.

2.1. DC Series Arc V-I Characteristics

This study simulates the series arc fault occurrence through an arc generator based on the UL1699B standard, consisting of two copper electrodes, insulators, a slide table, and a stepper motor. One of the copper rods with a flat tip is the stationary electrode, and the other with a pointed tip is the moving electrode, as shown in Figure 2.

The DC arc can be equivalent to a nonlinear resistance; the arc gap’s length and current will affect the arc voltage [37]. Once the series arc ignites, the arc voltage drastically increases to 15 V, as shown in Figure 3. The Nottingham model can represent the voltage and current characteristics of the DC series arc after arc stabilization, as shown in (1).

$$U_{\mathrm{arc}}=A+\frac{b}{I_{\mathrm{arc}}^a}$$

(1)

the values of A and b depend on the electrode material and the arc gap’s length, and the value of a depends on the electrode material.

According to the experiment, the DC arc inception voltage is approximately 15 V, minimally affected by the circuit current and the arc gap. The arc current determines whether the DC series arc can burn stably. If the arc current decreases to a certain value, the DC series arc will be extinguished.

2.2. Effect of the MPPT Algorithm under a Series Arc Condition

The PV module consists of a controlled current source, a resistance, a diode in parallel, and a series resistance as shown in Figure 4 [38]. According to Kirchhoff’s law, the U_PV-I_PV characteristics of the PV module can be expressed as follows:

$$I_{\mathrm{pv}}=I_{\mathrm{ph}}-I_{\mathrm{v}\mathrm{d}0}\left[\exp \left(\frac{U_{\mathrm{pv}}+I_{\mathrm{pv}}{R}_{\mathrm{s}}}{N_{\mathrm{s}}{U}_{\mathrm{t}}}\right)-1\right]-\frac{U_{\mathrm{pv}}+I_{\mathrm{pv}}{R}_{\mathrm{s}}}{R_{\mathrm{sh}}}$$

(2)

where I_ph is the photocurrent and its value depends on environmental factors, I_vd0 is the reverse saturation current of the diode, N_s is the number of the PV cells, R_s is the series resistance, U_t is the thermal voltage of the diode, R_sh is the parallel resistance of the PV module, and U_PV and I_PV are the PV output voltage and current.

The thermal voltage of the diode is represented as follows:

$$U_{\mathrm{t}}=\frac{KTA}{q}$$

(3)

where K is the Boltzmann constant, A is the diode quality factor, T is the temperature, and q is the electronic charge.

In practical engineering, the R_s is small, and the R_sh is large [39]. For a better analysis, Equation (2) can be simplified as:

$$I_{\mathrm{pv}}=I_{\mathrm{ph}}-I_{\mathrm{v}\mathrm{d}0}\left[\exp \left(\frac{U_{\mathrm{pv}}}{N_{\mathrm{s}}{U}_{\mathrm{t}}}\right)-1\right]$$

(4)

According to Equation (4), the PV output power characteristics can be obtained as:

$$P_{\mathrm{pv}}=U_{\mathrm{pv}}\left\{I_{\mathrm{ph}}-I_{\mathrm{v}\mathrm{d}0}\left[\exp \left(\frac{U_{\mathrm{pv}}}{N_{\mathrm{s}}{U}_{\mathrm{t}}}\right)-1\right]\right\}$$

(5)

According to Equation (5), the U_PV-I_PV characteristics of the PV array can be obtained to analyze the effect of series arc.

For a PV module under a specific environment, the PV output current and voltage are determined using its U_PV-I_PV characteristics and the load resistance of the PV module (R_PV). The relationship between the PV output and the R_PV under normal conditions is expressed as follows:

$$R_{\mathrm{PV}}=\frac{U_{\mathrm{pv}}}{I_{\mathrm{pv}}}=R_{\mathrm{eq}}+R_{\mathrm{line}}$$

(6)

where R_eq is the equivalent resistance of the inverter and R_line is the line resistance, which is relatively small and can be ignored.

The MPPT algorithm changes the R_eq by controlling the duty cycle D, thus achieving maximum power point (MPP) tracking [40].

$$R_{\mathrm{PV}}=R_{\mathrm{eq}}={(1-D)}^2{R}_{\mathrm{L}}$$

(7)

where R_L is the actual load impedance of the inverter, and D is the duty cycle of the MOSFET in the boost converter.

The perturb and observe method (P&O) is the common MPPT algorithm in PV systems. It searches the MPP voltage by comparing the power change before and after the disturbance period [41].

$$U_{\mathrm{mppt}}(T+1)=U_{\mathrm{mppt}}(T)+\Delta U$$

(8)

where U_mppt is the target voltage of the MPPT algorithm, and ΔU is the disturbance amplitude, which is positive or negative, indicating the direction of the disturbance. T is the disturbance period of the MPPT algorithm, and it is taken as 300 ms in this study.

Then, the MPPT algorithm changes the duty cycle D of the MOSFET via the PI controller to make the U_bus approach the U_mppt, as shown in Figure 5.

When the arcing process starts, the U_bus drastically decreases, but the U_mppt of the MPPT algorithm will not change due to the series arc. The PI controller compares the difference between the U_bus and the U_mppt per period. Then, the PI controller changes the duty cycle D based on the difference until the bus voltage equals the target voltage. This process is very short and depends on the dynamic speed of the PI controller. In this study, the period of the PI controller is taken as 10 ms.

Therefore, this study assumes that the U_bus remains almost unchanged at the MPP voltage before and after the series arc occurs due to the PI controller.

The U_bus is lower than the U_PV due to the series arc voltage. According to Kirchhoff’s law, the U_bus equals the difference between the U_PV and the U_arc.

$$U_{\mathrm{bus}}=U_{\mathrm{PV}}-U_{\mathrm{arc}}$$

(9)

2.3. PV Module Output Characteristics under a Series Arc Condition

Assuming that there are no changes in environmental factors before and after the series arc occurs, the U_PV-I_PV characteristics of the PV string remain unchanged before and after the series arc occurs. However, the faulty PV string voltage increases by the arc voltage compared to the U_bus. The faulty PV string U_bus-I_PV characteristics are as follows:

$$I_{\mathrm{pv}}=I_{\mathrm{ph}}-I_{\mathrm{v}\mathrm{d}0}\left[\exp \left(\frac{U_{\mathrm{pv}}-U_{\mathrm{arc}}}{N_{\mathrm{s}}{U}_{\mathrm{t}}}\right)-1\right]$$

(10)

Therefore, the U_bus-I_PV characteristics of the faulty PV string deviate to the left compared to the U_PV-I_PV characteristics. The amount of left shift is the magnitude of the series arc voltage. Furthermore, the U_bus-I_PV characteristics of the normal PV string are equal to the U_PV-I_PV characteristics. This study assumes a constant arc voltage of 15 V for better analysis, as shown in Figure 6.

In the figure, U_mpp and U_oc arc the PV voltage at the MPP and open circuit point under normal conditions. U_arc,mpp and U_arc,oc are the U_bus at the MPP point and open circuit point under series arc conditions.

The PV voltage increase due to the series arc causes the PV operating point to move toward the open circuit point. Then, the PV output current decreases, and the decreased value is related to the parameters of the PV module.

$$\Delta I=I_{\mathrm{v}\mathrm{d}0}\left[\exp \left(\frac{U_{\mathrm{pv}}}{N_{\mathrm{s}}{U}_{\mathrm{t}}}\right)-\exp \left(\frac{U_{\mathrm{pv}}+U_{\mathrm{arc}}}{N_{\mathrm{s}}{U}_{\mathrm{t}}}\right)\right]$$

(11)

The parameters of the PV module will not change due to the series arc, but these parameters are difficult to calculate accurately in practical engineering. Therefore, this study estimates the current drop ΔI based on the PV output characteristic curve:

$$\Delta I=\frac{k{U}_{\mathrm{arc}}{I}_{\mathrm{loop}}}{(U_{\mathrm{oc}}-U_{\mathrm{mpp}})}$$

(12)

The parameter k depends on the curvature of the PV characteristic curve at the MPP point. According to the experiment and simulation, the range of k values is 0.5~0.7; it depends on PV parameters and aging degree. As the PV module ages, the series resistance R_s gradually increases, the U_mpp gradually decreases, and the U_oc remains unchanged. Therefore, the current drop value decreases under the series arc fault condition with PV module aging. In this study, k is taken as 0.5.

If the series arc occurs on one of the strings, then that string is the faulty string whose current decreases by ΔI, and the other PV strings are normal strings whose current remains unchanged. If the series arc occurs on the main PV bus, all PV strings are faulty strings whose current decreases by ΔI.

Whether the series arc fault can maintain stable combustion is closely related to the voltage of the series arc, the MPPT algorithm, and the U_PV. If the series arc voltage is equal to the difference between U_oc and U_mpp, when the series arc occurs, the U_bus remains almost unchanged due to the MPPT algorithm, and the U_PV increases to the U_oc. Then the PV output current almost drops to zero, and the series arc will extinguish, resulting in a faulty string open circuit. Similarly, the series arc can be extinguished by increasing the bus voltage.

The weakening of light intensity also causes a decrease in the PV output current, which slowly decreases under this condition. However, under the series arc condition, the PV output current will steeply drop due to the series arc voltage drastically increasing when the series arc occurs. This can be used as a feature to distinguish between the light variation condition and the series arc. Another feature is that the PV system does not open-circuit due to a low current under normal conditions.

3. The Series Arc Detection Method for the PV System

The loop current will steeply drop due to the series arc voltage drastically increasing when the series arc occurs. This study detects the series arc through the current drops by the current rate change, as follows (13).

$$I_{\max }(t)-I_{\min }(t)\ge I_{\mathrm{set}}$$

(13)

where I_set is set for detecting the current drop due to the series arc. The current drop ΔI can be estimated using Equation (12). Considering the influence of the perturbation, the detection threshold I_set takes 0.8 times the current drop ΔI. t is the detection period, taking into account the dynamic response time of the PI controller. The period of 50 ms is chosen. I_min(t) and I_max(t) are the minimum and maximum values within the detection period.

Generally, the variation in light intensity within a period of 50 ms does not cause a drastic change in the PV output current. However, sudden shading may cause the PV output current to drop steeply, like the series arc fault. To distinguish this situation and extinguish the series arc, this study suppresses series arc combustion by increasing the bus voltage to the U_oc that is under the series arc fault.

The PV module belongs to nonlinear power sources. If the PV voltage increases to the U_oc, the PV output current will drop to zero. Under normal conditions, the U_oc of the PV string is almost equal to the U_oc of the bus. Under the series arc fault, the open circuit voltage of the bus (U_arc,oc) decreases due to an extra arc voltage insertion, as follows:

$$U_{\mathrm{arc},\mathrm{oc}}=U_{\mathrm{oc}}-U_{\mathrm{arc}}$$

(14)

The U_oc is approximately 1.2~1.3 times the U_mpp, and the minimum voltage of the DC series arc is about 15 V. Therefore, U_arc,oc can be estimated by the U_mpp that is the PV voltage before the occurrence of the series arc. If the bus voltage increases the U_arc,oc, the fault point will open-circuit due to the arc extinguishing under a low current. Then, the current of the loop where the series arc is located will drop to zero.

In summary, the proposed method utilizes the drop of the loop current to detect the series arc and extinguishes the series arc by increasing the target voltage to the U_arc,oc. This is shown in Figure 7, which illustrates the fault diagnosis process.

4. Test Verification

In order to test and verify the practicality of the proposed method, a boost converter is built and tested. The MPPT algorithm and the series arc detection method were implemented in the STM32f103 microcontroller. The MPPT algorithm is the P&O method with a fixed step size.

The PV experimental platform is shown in Figure 8. The PV module parameters are shown in

Table 1. Parameters of the PV system.

	Parameters	Values
PV module	Pmax	269.98 W
	Vmax	30.68 V
	Imax	8.80 A
	Voc	38.39 V
	Isc	9.29 A
MPPT algorithm	Disturbance period	300 ms
	PI control period	10 ms
	Step size	1 V

. The cable and conductor are used to satisfy the line impedance requirements of the UL1699B standard.

The arc generator is placed in different locations in the experimental platform: on the main PV bus, the PV string, and between the PV modules.

Three PV modules form a PV system with one string and the arc generator was placed on the main PV bus. Under the series arc condition, the waveform of the input voltage (bus voltage) and current (loop current) of the converter are shown in Figure 9.

Under the converter without the proposed method condition, the waveform is as shown in Figure 9a. Stage Ⅰ (0–4.1 s) operates normally, and the bus voltage and loop current oscillate around the MPP due to the P&O method with a fixed step size. At 4.1 s, the series arc occurs. The bus voltage remains unchanged due to the PI controller, and the loop current drops to 3.5 A. In Stage Ⅱ (4.1–10.1 s), the MPPT algorithm increases the PV output power through disturbance. Stage Ⅲ (10.1–20 s) is influenced using the MPPT algorithm, and the PV system reaches the new MPP. Then, the bus voltage decreases by the arc voltage, and the loop current returns to the level before the occurrence of the series arc.

Figure 9a shows that the PV series arc can stably burn without the proposed method. In the PV system, the loop current has no zero-crossing point. Hence, once the series arc occurs in the main PV bus, it can persist, decreasing PV output power and even posing fire hazards.

Under the converter with the proposed method condition, the waveform is shown in Figure 9b. Stage Ⅰ (0–4 s) operates normally, and the bus voltage and loop current oscillate around the MPP. At 4 s, the main PV bus causes the series arc fault, causing the loop current to drop by 3A. Moreover, the change in the current rate exceeds the threshold. In Stage Ⅱ (4–4.5 s), the series arc detection algorithm increases the target voltage of the MPPT algorithm in the next disturbance cycle. The bus voltage increases, and the loop current begins to decrease. In Stage Ⅲ (4.5–5 s), the series arc is extinguished when the loop current decreases to 1.5 A, and the position of the series arc opens circuit. Then, the bus voltage and the loop current drop to zero.

Figure 9b shows that the proposed method can successfully detect and extinguish the series arc that occurs on the main PV bus. Furthermore, it took approximately 0.5 s from arc generation to when the arc was extinguished, and the PV bus link open circuit due to the series arc was extinguished.

After increasing the number of PV modules to 5, and placing the arc generator on the main PV bus, the test waveform is shown in the Figure 10.

As shown in Figure 10, the bus voltage increases, and the loop current remains unchanged. Stage Ⅰ (0–2.5 s) operates normally, and the bus voltage and loop current oscillate around the MPP. At 2.5 s, the series arc occurs, and the loop current drastically decreases by 1.5 A. In Stage Ⅱ (2.5–2.9 s), the proposed method increases the target voltage of the MPPT algorithm in the next disturbance cycle. The bus voltage increases and the loop current begins to decrease. In Stage Ⅲ (2.9–5 s), the series arc is extinguished when the loop current decreases to 1.5 A, and the position of the series arc opens circuit. Then, the bus voltage and the loop current drop to zero due to the bus link open circuit.

As the number of PV modules increases, the value of the current drop decreases, and the additional open-circuit voltage required increases. Then, the proposed method still successfully detects and extinguishes the series arc within 0.5 s.

In order to test and verify the practicality of the proposed method for the series arc occurring in the PV string, six PV modules were formed into a PV array (three PV modules in series and two strings in parallel), and the arc generator was placed in the PV strings.

Under the converter without the proposed method condition, the waveform is shown in Figure 11.

Stage Ⅰ (0–3.9 s) operates normally, and the bus voltage and loop current oscillate around MPP. At 3.9 s, the series arc occurs, and the loop current drastically decreases by 3A. In Stage Ⅱ (3.9–6.1 s), the MPPT algorithm increases the PV output power through disturbance. In Stage Ⅲ (6.1–20 s), with the influence of the MPPT algorithm, the PV system reaches the new MPP, and the loop current returns to the level before the fault occurs.

The series arc can stably burn in the PV string without the proposed method. Under the series arc located at the PV string condition, the faulty string MPP voltage decreases, but the normal string MPP voltage remains unchanged. Under the series arc located at the main PV bus condition, the MPP voltage of all PV strings decreases. Therefore, the new MPP voltage at which the series arc occurs on the PV string is higher than the new MPP voltage at which the series arc occurs on the main PV bus.

After placing the arc generator on the PV string and between the PV module as positions #2 and #3 in Figure 1, under the converter with the proposed method condition, the waveform is shown in Figure 12.

As shown in Figure 12, the impacts of the series arc occurring in the PV string and between the PV modules within the string on the PV system are the same. Stage Ⅰ (0–1.1 s) operates normally. At 1.1 s, the series arc occurs. The loop current decreases by 3 A. In Stage Ⅱ (1.1–1.9 s), the proposed method increases the target voltage of the MPPT algorithm. Then the bus voltage increases, and the loop current decreases. The series arc extinguishes due to a low arc current during the increase in bus voltage. Then, the PV faulty string output current steeply drops to zero. In Stage Ⅲ (1.9–7 s), the MPPT algorithm increases the PV output power through disturbance, and the series arc fault will not reignite. In Stage Ⅳ (7–9 s), the MPPT algorithm is reworked, and the PV system reaches the new MPP.

The results show that the proposed method can successfully detect and extinguish the series arc that occurs in the PV string. The PV faulty string opened circuit after the series arc extinguishing. Another string normal operation and the series arc will not reignite in a low voltage. Therefore, when the PV system reaches the new MPP, the loop current is only half the value before the series arc fault, and the bus voltage remains unchanged.

Testing the series arc detection algorithms under light variation conditions in cloudy weather conditions, the current waveform is shown in Figure 13.

As shown in Figure 13, the loop current under light variation conditions does not drastically decrease like the series arc fault condition. Therefore, the proposed method was verified to be immune to light variation under cloudy weather conditions.

In summary, the proposed method can successfully detect and extinguish the series arc without additional hardware and complex algorithms in the PV system, and the false detection due to light variation under cloudy weather conditions did not occur.

5. Conclusions

The main conclusions of this study are as follows:

(1)

Once the series arc occurs, the input voltage of the converter remains unchanged due to the MPPT algorithm, and the PV output current will drastically decrease. On this basis, a series arc detection method based on the current change was proposed;

(2)

The DC series arc is difficult to burn stably under a low arc current. On this basis, a series arc extinguishing method based on increasing the target voltage of the MPPT algorithm was proposed;

(3)

Compared to the traditional series arc detection methods, the proposed methods can be integrated into the PV system without additional hardware. The results indicate that the proposed method can successfully detect and extinguish the series arc in the low-voltage PV system.