Effects of Welding Time and Electrical Power on Thermal Characteristics of Welding Spatter for Fire Risk Analysis

Abstract

To predict the fire risk of spatter generated during shielded metal arc welding, the thermal characteristics of welding spatter were analyzed according to different welding times and electrical powers supplied to the electrode. An experimental apparatus for controlling the contact angle between the electrode and base metal as well as the feed rate was prepared. Moreover, the correlations among the volume, maximum diameter, scattering velocity, maximum number, and maximum temperature of the welding spatter were derived using welding power from 984–2067 W and welding times of 30 s, 50 s, and 70 s. It was found that the volume, maximum diameter, and maximum number of welding spatters increased proportionally as the welding time and electrical power increased, but the scattering velocity decreased as the particle diameter increased regardless of the welding time and electrical power. When the measured maximum temperature of the welding spatter was compared with an empirical formula, the accuracy of the results was confirmed to be within ±7% of the experimental constant $C=112.414\times P_e^{-0.5045}$. Results of this study indicate quantitatively predicting the thermal characteristics of welding spatter is possible for minimizing the risk of fire spread when the electrode type and welding power is known.

1. Introduction

Fire risks in construction sites may occur when flammable gases, liquids, or substances reach their ignition points owing to the scattered welding spatter [1,2,3,4,5,6]. Shielded metal arc welding (SMAW), a method of joining metals by generating an arc and heating the weld metal zone by applying electrical power between the base metal and electrode, has been widely used in industrial sites since the method of SMAW is applied to almost all repairing of cast iron in air or steel under water [7,8,9,10]. However, it involves the risk of fire spreading to nearby combustibles caused by high temperatures because the scattered welding spatters are larger comparable to those of gas metal arc welding (GMAW), which uses plasma [11,12]. Especially, the fire hazards from the SMAW at building construction sites can occur when welding spatters make contact with the inward of a pipe or other enclosed space filled with flammable vapor or liquid [12,13,14,15,16,17,18,19]. In addition, the polarity of electrode can cause changes in welding spatter diameter and number [7,19,20,21,22,23]. From the viewpoint of fire technology, analyzing the thermal characteristics of welding spatter is one of the widely used methods to predict fire spread, where related research has already been conducted.

Hagiwara et al. [24,25] conducted an experimental study on the particle size distribution of welding spatter according to the electrical power supplied to the electrode. They found that 90% of the particles had diameters of less than 1 mm and analyzed the fire spread phenomenon in combustibles, such as benzene, acetone, and urethane foam. This study, however, appears to have limitations in quantitatively analyzing the thermal characteristics of welding spatter crucial to fire risks, which are caused by the electrical power and depend on the particle size.

Hagimoto et al. [19] calculated the particle size according to the electrode diameter when the same electrical power was supplied to the electrode and found that approximately 80% of the particles were scattered to a distance of 0.5 m, 15% to 0.5–1.0 m, and 5% to more than 1 m. They reported that fire can spread to combustibles (urethane foam etc.) when large particles of diameters 0.9–3.0 mm are scattered to a distance of more than 3.5 m.

Brandi et al. [26] analyzed the correlation between the material properties of the electrode core and fire risks using standard mineral dressing techniques. They found that the porosity and density of the welding spatter varied according to the electrical power and stressed the importance of the electrode physical properties for satisfying the ignition requirements of combustibles.

Results from previous studies show that the conditions of fire spread to combustibles during welding vary due to the varying thermal characteristics of welding spatter depending on the electrical power [18,26,27]. Therefore, Shin and You [27] calculated the particle size distribution and mean particle size of welding spatter by assuming a steady-state maximum temperature of the welding spatter for igniting combustibles and proposed an equation for predicting the mean particle temperature based on the energy conservation relationship. According to them, predicting the maximum temperature of the welding spatter is possible when the electrical power, total volume, mean size, and scattering velocity of the particles are known. As these parameters (except the electrical power) vary depending on the electrical power, it is necessary to analyze the relationships among the main factors according to the experimental conditions. This is necessary for the quantitative analysis of the risk of fire spread due to scattered particles. Therefore, in this study, we propose a method for predicting fire risks by quantitatively deriving the thermal characteristics of welding spatter according to the welding time and electrical power.

2. Material and Methods

2.1. Theoretical Approach

Figure 1 shows the schematic of the total volume of the welding spatter during welding. The total mass of the welding spatter (Δm_p,total) can be calculated according to Equation (1) after measuring the mass melted on the base metal (Δm_b,p) and is dependent on the welding time (Δt) and electrical power (P_e). The core inside the electrode is made of steel (ρ_iron = 7860 kg/m³) and the coating outside the electrode contains sodium silicate (ρ_{Sodium Silicate} = 2400 kg/m³). However, it is possible to calculate the volume of a single particle (ΔV_i) using the relationship ΔV_i = Δm_i/ρ_i only when the mixed ratios of materials are given for each welding spatter.

$$\Delta m_{p,total}=\Delta m_{el}-\Delta m_{b,p}$$

(1)

where $\Delta m_{el}$, $\Delta m_{b,p}$, and $\Delta m_{p,total}$ are the masses of the electrode and solidified weld metal attached to the base metal and total mass of scattered particles, respectively. In a previous study, the mean particle temperature was predicted by assuming the steady-state condition of the initial temperature of welding spatter as the maximum value for the ignition combustibles, as shown in Equation (2) [27].

$$T_{P,s}=T_{\infty }+\frac{P_e-\sigma \epsilon A_{b,s}\left(T_{s,b}^4-T_{sur}^4\right)}{Nh{A}_{P,s}}$$

(2)

where $T_{p,s}$, $T_{\infty }$, $T_{sur}$, $P_e$, $\sigma$, $\epsilon$, $A_{b,s}$, $T_{s,b}$, N, h, and $A_{p,s}$ are the mean particle temperature, surrounding temperature, surface temperature, electrical power (P_e), Stefan–Boltzmann constant, emissivity, surface area and surface temperature of base metal, average number of particles, convective heat transfer coefficient, and surface area of particle, respectively. The mean particle size, $d_{p,m}$, and convective heat transfer coefficient, h, can be obtained using Equations (3) and (4), respectively [14,27,28].

$$N=\frac{6V_{p,total}}{\pi d_{p,m}^3}$$

(3)

$$N{u}_D=\frac{h{d}_{p,m}}{k}=2+0.6{Re}_D^{0.5}P{r}^{1/3}$$

(4)

where V_p,total, d_p,m, $N{u}_D$, $k$, Re, and Pr are the total volume of particles, mean diameter of particles, Nusselt number, thermal conductivity, Reynolds number (Re_D = ρu_p,md_p,m/μ), and Prandtl number (Pr = C_pμ/k)), respectively. Therefore, the temperature distribution prediction shown in Equation (2) is possible when V_p,total, d_p,m, and u_p,m can be calculated using Equations (3) and (4). As the total volume, mean diameter, and scattering velocity of particles vary according to the welding time and electrical power, the functional relationship given by Equation (5) must be also determined [14,27].

$$N,d_m,h~f(P_e,\Delta t)$$

(5)

2.2. Experimental Apparatus

Figure 2 shows the semi-automated SMAW experimental apparatus, which was constructed by maintaining perpendicularity between the electrode and base metal constant (welding angle θ = 90°) and specifying the maximum feed rate of the welding torch as 7 mm/s. This made it possible to analyze the size and scattering velocity of the particles. As shown in the figure, welding spatters were scattered under different welding times (Δt) and electrical power (P_e), and the scattering velocity and mean temperature of the welding spatters were measured using a high-speed camera (model: phantom Miro M/R/LC310, USA) and a thermal imaging camera (model: Fluke Tix501). The scattering velocity and mean temperature of the welding spatter were measured using a high-speed camera (model: phantom LC310) and a thermal imaging camera (model: Fluke Tix501). The particle size distribution was determined using Image J software after collecting all welding spatter in a 50 × 46 × 64 cm³ acrylic box.

Table 1. Specifications of experiment apparatus.

Equipment	Specification
Welding machine	Output current (20~220) A, Rated input voltage 220 V, Electric power (0~2.5) kW Rated duty cycle 60%, Model: Rolwal MMA-200E
Thermal imaging camera	Infrared resolution 640 × 480, Temp. measurement range −20 to 650 °C, Accuracy ±2 °C or 2%, Frame rate 60 Hz, Model: Fluke Tix 501
High-speed camera	Resolution 640 × 480, Sampling rate 10,000 fps, Model: phantom Miro M/R/LC310 Lens: Nikon 105 mm, 2× converter Band-pass filter: Φ 50 mm, 810 nm/12 nm, CN code 90022000
Electronic energy meter	230 AC, 60 Hz, 16 A/3680 W (Model: KEM2500)
Precision balance	Max. load weight 320 g, Accuracy 0.1 mg, Model: PX224KR
Electrode	High titanium oxide type electrode (AWS E-6013) Core: Iron (65–75%), Coating: Titanium dioxide (10–15%), Feldspar (5–10%), Mn (1–5%), Sodium silicate (1–5%), Limestone (1–5%), Mica (1–5%)

lists the specifications of the experimental apparatus and the experimental conditions for the average values of three times results are denoted in

Table 2. Experimental conditions to study the effects of thermal characteristics of welding spatter on the welding time and electrical power.

Test Number	Welding Time, Δt (s)	Welding Current (A)	Welding Voltage (V)	Electrical Power, Pe (W)
Case #1	30	80	12	984
Case #2	50
Case #3	70
Case #4	30	100	13	1337.3
Case #5	50
Case #6	70
Case #7	30	130	14	1802.0
Case #8	50
Case #9	70
Case #10	30	150	17	2067.5
Case #11	50
Case #12	70

Welding polarity: DC-, Contact angle: 90°, Arc length: 5 mm, Base metal: Mild steel (SS400); Electrode diameter, d_el: 4.0 mm, Material properties of electrode given in Table 1.

.

3. Results and Discussion

3.1. Volume of Welding Spatter

Figure 3 shows the variation of the measured reduction rate (u_rate) of the electrode length according to the electrical power (P_e) in the case of welding times (Δt) of 30 s, 50 s, and 70 s. It is seen that as P_e increased, u_rate also increased proportionally as the mass of the electrode welded to the base metal (Δm_b,p) increased. When P_e was constant, a constant value of u_rate was calculated, which was consistent with that obtained using Equation (6) within ±4% for the average values u_rate regardless of Δt.

$$u_{rate}=a_1+b_1\times P_e$$

(6)

Here, a₁ and b₁ are experimental constants. It is estimated that a₁ = 0.989 mm/s and b₁ = 0.157×10⁻² W-mm/s are obtained according to the base metal and electrode specifications listed in Table 1, and the electrical power ranges between 984 and 2067 W.

Figure 4 shows the variation in the measured mass loss of the electrode (Δm_el) and the mass welded to the base metal (Δm_b,p) according to the electrical power (P_e) for the welding times (Δt) of 30 s, 50 s, and 70 s. In this figure, the symbols enclosed in brackets represent Δm_el, which increased proportionally to u_rate. When $\Delta t$ increased keeping P_e constant, Δm_el increased in proportion to $u_{rate}$. Therefore, when Equation (6) and the electrode density measured using a load cell (ρ = 4726 kg/m³) are applied, Δm_el is given by Equation (7) expressed by the dotted line, which agrees with the measured value within an error range of approximately ±5%.

$$\Delta m_{el}=u_{rate}\times \Delta t\times \left(\frac{\pi }{4}{d}_{el}^2\right)\times {\rho }_{el}$$

(7)

where d_el is the electrode diameter is used as the reference value of 4.0 mm. Notably in the figure, the measured value of Δm_b,p increased in proportion to the magnitudes of Δt and P_e, as shown by the closed symbol value. In addition, 88.6% Δm_el was found to be welded to the base metal on average. This result indicates that approximately 11.4% Δm_el was responsible for generating welding spatter when P_e was supplied to the electrode. The energy transmitted to the electrode can be simplified through the assumption shown in Equation (8).

$$\sigma \epsilon A_{b,s}\left(T_{s,b}^4-T_{sur}^4\right)\equiv 0.886P_e$$

(8)

Figure 5a shows the variation of the total mass of the particles scattered from the electrode (Δm_p,total) using Equation (1), mass reduction of the electrode (Δm_el), and mass welded to the base metal (Δm_b,p) for the welding times (Δt) of 30 s, 50 s, and 70 s according to the electrical power. The result of curve-fitting the calculated values of Δm_p,total with increasing electrical power (P_e) under the same Δt values is shown in Equation (9).

$$m_{p,total}=a_2\cdot P_e^{b_2}$$

(9)

It was found that a₂ is related to $\Delta t$ as shown in Figure 5b, and this tendency is shown in Equation (10) when b₂ is constant at 1.28944.

$$a_2=2.9\times {10}^{-5}\mathrm{g}+1.23\times {10}^{-6}\mathrm{g}/\mathrm{s}\times \Delta t$$

(10)

Figure 6 shows the density values (ρ_i) of a single scattered welding particle measured by calculating the mass (m_p,i) and volume (ΔV_p,i) of the particle using diameters (d_p,i) of 1.736, 2.023, 2.294, and 2.352 mm. Because the electrode contains various metal components, as shown in Table 1, ρ_i may vary depending on the material composition inside the shield and core [26,29]. In particular, the mass proportions of metals that constitute each particle must be determined to obtain the total volume of scattered welding spatter (ΔV_p,total), but limitations exist in analyzing the density when measuring each mixed component for at least 1000 small particles with a diameter of 0.1 mm or less. Therefore, we attempted to analyze the thermal characteristics of welding spatter by assuming ρ_p,total ≡ ρ_el (4726 kg/m³) and calculating ΔV_p,total as shown in Equation (11).

$$\Delta V_{p,total}=\frac{\Delta m_{p,total}}{{\rho }_{el}}=\frac{(2.9\times {10}^{-5}+1.23\times {10}^{-6}\times \Delta t)\times P_e{}^{b_2}}{{\rho }_{el}}$$

(11)

3.2. Diameter and Number of Welding Spatters

Figure 7 shows the fraction (N_i/N_total) of the number of particles, which is the ratio of particles with diameter (N_i) to the total number of scattered particles (N_total), according to P_e at Δt = 30 s, 50 s, and 70 s. As mentioned before, the particle size distribution was determined using Image J software after collecting welding spatter in a 50 × 46 × 64 cm³ acrylic space, and it was analyzed by excluding the diameters of 0.3 mm or less due to the resolution. It was found that the mean particle diameter (d_p,m) was approximately 0.3 mm regardless of $\Delta t$ and P_e, which is similar to the results of previous studies (d_p,m < 0.5 mm) [9,11]. Therefore, the mean number of particles (N) can be calculated using Equation (4). The main purpose of this study was to predict fire risks according to the thermal characteristics of scattered particles; however, the mean number of particles (N) was expressed using the maximum number of particles (N_max) to analyze the thermal characteristics according to the maximum particle diameter (d_p,max) generated during welding.

As the maximum particle size, d_p,max, is still undetermined, it is necessary to analyze d_p,max according to Δt and P_e to solve Equation (12).

$$N_{max}=\frac{6V_{total}}{\pi d_{p,max}^3},N=N_{max}{\left(d_{p,max}/d_{p,m}\right)}^3$$

(12)

Figure 8 shows the results of analyzing the maximum diameter of scattered particles (d_p,max) according to the electrical power (P_e) at Δt = 30 s, 50 s, and 70 s. Each measured value represents the average of the maximum diameter obtained in three repeated experiments. Apparently, the size of the particles scattered from the electrode increased as Δt increased under a constant P_e because the temperature around the weld zone of the base metal increased. In addition, under the same Δt, the size of scattered particles increased in proportion to the melted mass of the electrode as P_e increased as shown in Equation (13).

$$d_{p,max}=a_3\times P_e{}^{b_3}$$

(13)

where a₃ and b₃ are experimental constants. When b₃ = 0.4825, a₃ can be calculated by Equation (14) and plotted in Figure 8b.

$$a_3=2.194\times {10}^{-2}+9.38\times {10}^{-4}\times \Delta t$$

(14)

3.3. Velocity of Welding Spatter

Figure 9a shows the results of measuring the mean particle velocity according to the particle diameter under an electrical power of 984 W using a high-speed camera (model: phantom LC310) and particle tracking velocimetry (PTV). It is observed that the scattering velocity showed a tendency to decrease as the particle diameter increased under the experimental conditions of the welding time (Δt) and electrical power (P_e), as shown in Table 2. The correlation between the maximum diameter of scattered particles (d_p,max) and the scattering velocity (u_p,max) was analyzed, as shown in Figure 9b. u_p,max decreased as d_p,max increased with a difference of less than ±10% depending on the values of Δt and P_e, and the relationship shown in Equation (15) was found.

$$u_{p,max}=1.3\times d_{p,max}^{-1.004}$$

(15)

3.4. Thermal Characteristics of Welding Spatter

Using the results of the total volume of particles (V_p,total), maximum number of particles (N_max), particle diameter (d_p,max), and scattering velocity (u_p,max) according to the Δt and P_e obtained in Section 3.1 to determine the convective heat transfer coefficient shown in Equation (4), Equation (16) can be formed.

$$h_{max}=\left(2+0.6{Re}_{d,max}^{0.5}P{r}^{1/3}\right)k(T_{ref})/d_{p,max},$$

(16)

where Re_d,max is the Reynolds number (Re_d,max = ρ_pu_p,maxd_p,max/μ) considering the maximum particle diameter (d_p,max). In particular, the thermal conductivity (k), specific heat (Cp), viscosity (μ), and density (ρ) of air vary depending on the reference temperature (T_ref = (T_p,s + T_∞)/2), as shown in Figure 10a. In this study, these can be calculated using Equations (17)–(20) based on the data given by this study [30].

$$k(T_{rep})=-7.16\times {10}^{-3}+1.72\times {10}^{-4}\times T-2.45\times {10}^{-7}\times T^2+2.29\times {10}^{-10}\times T^3-9.81\times {10}^{-14}\times T^4+1.63\times {10}^{-17}\times T^5$$

(17)

$$c_p(T_{rep})=1.05-2.89\times {10}^{-4}\times T+6.83\times {10}^{-7}\times T^2-3.4\times {10}^{-10}\times T^3+2.25\times {10}^{-14}\times T^4+1.55\times {10}^{-17}\times T^5$$

(18)

$$\mu (T_{rep})=4.26\times {10}^{-6}+4.93\times {10}^{-8}\times T+-1.33\times {10}^{-11}\times T^2+2.36\times {10}^{-15}\times T^3$$

(19)

$$\rho (T_{rep})=374.074\times T^{-1.0114}$$

(20)

Figure 10a shows the results of analyzing the convective heat transfer coefficient, h, according to T_ref when d_p,max = 0.1, 0.3, 1.0, and 3.0 mm. h decreased as d_p,max increased while the scattering velocity (u_p,max) decreased to 13.12, 4.35, 1.30, and 0.43 m/s at different d_p,max values obtained by Equation (15). Therefore, Equation (2), for calculating the mean particle temperature considering the welding time and electrical power, can be expressed as in Equation (21).

$$T_{P,s}=T_{\infty }+\frac{0.13P_e}{N_{max}{h}_{max}{A}_{P,max}{r}_{ratio}},$$

(21)

where N_max, h_max, A_p,max, and r_ratio are the total number of particles, heat transfer coefficient, surface area of a particle, and a constant calculated by replacing d_p,m with d_p,max, respectively, when welding particles are at their maximum size. In particular, the mean number of particles (N) used in Equation (3) and the mean surface area (A_p,m) are related as N = N_max(d__p,m/d__p,max)⁻³ and A_p,m = A_p,max(d__p,m/d__p,max)², whereas the convective heat transfer coefficient (h) for determining the temperature of the welding spatter is given by h = h_max × (d__p,m/d__p,max)^−0.5 × (u_p,m/u_p,max) ^0.5. Therefore, r_ratio is related as,

$$r_{ratio}~{\left(\frac{u_{p,m}}{u_{p,max}}\right)}^{0.5}{\left(\frac{d_{p,m}}{d_{p,max}}\right)}^{-1.5},$$

(22)

where, u_p,m/u_p,max represents the ratio of the mean velocity to the maximum velocity. As it varies depending on the diameter, it can be expressed as Equation (23).

$${\left(\frac{u_{p,m}}{u_{p,max}}\right)}^{0.5}=C{\left(\frac{d_{p,m}}{d_{p,max}}\right)}^{0.5},$$

(23)

where C is the experimental constant which may vary depending on the velocity difference. Because k, Cp, μ, and ρ used to obtain the convective heat transfer coefficient are functions of T_p,s as shown in Equations (17)–(20), it is necessary to solve Equation (13) for the maximum particle size, Equation (15) for the maximum velocity, Equation (16) for the convective heat transfer coefficient, and Equation (23) to perform iterative calculations for predicting the maximum temperature of the welding spatter.

Figure 11a shows the results of maximum temperature (T_p,max) measured by capturing the welding spatter images accumulated on the acrylic collection plate at 60 fps for 70 s using a thermal imaging camera (Model: Fluck Ti520) at Δt = 70 s and P_e = 1337 W. As shown in the figure, the approximate maximum and mean particles were 432 °C and 347 °C, respectively.

Figure 11b shows the results calculated using the maximum particle temperature measurements and Equation (21) under the experimental conditions of Δt and P_e shown in Table 2. The experimental values agreed with the mean values with a difference of up to ±10% depending on Δt, and the temperature tended to increase as P_e increased. It should be noted that the maximum difference due to the time change was small (within ±2C) as shown in Equation (21) when C was constant, but the increasing tendency of the temperature in proportion to P_e was found to be consistent with the experimental values. However, when the values of C were 1, 2, 3, and 4, the slope at which the maximum particle temperature increased over the increase in P_e was smaller than the experimental value. This appears to be due to different values of C when the difference between the mean and maximum scattering velocities of the welding spatter increased along with P_e. At each P_e, the value of C can be obtained using Equations (24) and (25).

$$C_1=309.67\times P_e^{-0.6876}$$

(24)

$$C_2=112.414\times P_e^{-0.5045}$$

(25)

C₁ is a constant calculated by performing iterative calculations using Equations (17)–(20) for predicting the maximum particle temperature, whereas C₂ is calculated using the values of the density, thermal conductivity, viscosity coefficient, and specific heat at room temperature (298 K). Therefore, the maximum particle temperature can be predicted within an error range of approximately 5% using the equation to solve C₂.

Figure 12 shows the results of calculating the maximum temperature and diameter of the welding particles when the welding time (Δt) ranged from 30–70 s and P_e from 984–2067.5 W. “Fire hazard region” means the possibility of fire spreading to combustible materials such as polyurethan foam as mentioned in Ref [14,15,27], and “No ignition region” means the minimized conditions of fire spread. Based on previous studies, it can be confirmed that the maximum welding time and electrical power are 10 s and 1150 W, respectively, when the minimum particle size of welding spatter for the risk of fire spread is 0.9 mm, and the minimum temperature is 350 °C [19]. Therefore, the results of this study indicate that it is possible to calculate the electrical power for minimizing the risk of fire due to welding spatter when the electrode type and welding time are known.

4. Summary

In this study, the volume, maximum diameter, scattering velocity, and maximum number of welding spatter for shielded metal arc welding (SMAW) were analyzed according to the electrical power and welding time. When the electrical power was varied for welding times of 30 s, 50 s, and 70 s, the following results were derived.

First, when the mass of the electrode and scattered particles was calculated, an empirical formula was derived, which showed an increase in the mass of scattered particles when the electrical power increased at a constant welding time. In particular, the mass of scattered welding spatter represented approximately 11.45% of the total mass of the consumed electrode on average. The densities of the scattered particles were found to vary between 4876–7572 kg/m³ depending on the volume fraction of the core and coating composition of the electrode as referred by manufacturer.

Second, it was found that the mean diameter of welding spatter was approximately 0.3 mm, which was constant regardless of the welding time and electrical power. The maximum particle size, which has an important impact on fire risks, however, showed a tendency to increase in proportion to the welding time and electrical power. An empirical formula considering the maximum particle size was also derived to predict the temperature of the scattered welding spatter.

Third, the scattering velocity differed with differences of up to ±91% according to the welding time and electrical power. This appears to be due to the fact that materials with significantly different densities were mixed, which affected the momentum of the welding spatter while they were generated from the electrode. However, the scattering velocity decreased as the particle diameter increased.

Fourth, empirical formulas for the volume, maximum diameter, and scattering velocity of the welding spatter according to the welding time and electrical power were derived and compared with the maximum temperature measurements during the welding process. Results showed a good agreement between the compared values within an error range of approximately 10%. After verifying this accuracy, the case in which the minimum temperature of welding spatter was 350 °C or higher and the particle size was 0.9 mm was analyzed. It was found that fire risks can be minimized when a maximum welding time of 10 s and maximum electrical power of 1150 W are used. It should be noted that the maximum temperature of the welding spatter increased in proportion to the electrical power regardless of the welding time. The results of this study are expected to be used as important data for quantitatively presenting measures to minimize the fire risk of welding spatter.