Arc Discharge–Induced Ignition of Combustibles Placed on a Damaged AC Power Supply Cord

energies

Article Arc Discharge–Induced Ignition of Combustibles Placed on a Damaged AC Power Supply Cord

Kiyoto Takenaka 1,2,* , Yusuke Ishikawa 1, Yukio Mizuno 1,* and Wenyi Lin 3

1 Department of Electrical and Mechanical Engineering, Nagoya Institute of Technology, Nagoya 466-8555, Japan; [email protected] 2 Industrial Research Center, Aichi Center for Industry and Science Technology, Kariya 448-0013, Japan 3 Research & Development Division, Kawamura Electric Inc., Seto 489-8611, Japan; [email protected] * Correspondence: [email protected] (K.T.); [email protected] (Y.M.); Tel.: +81-566-24-1841 (K.T.); +81-52-735-5439 (Y.M.) Received: 19 December 2019; Accepted: 30 January 2020; Published: 5 February 2020

Abstract: One of the major causes of unintentional electrical fires is short circuit of an electrically and/or mechanically damaged alternating power supply cord. Detecting of such an event and interrupting the power supply may be beyond the capability of a conventional electro-mechanical circuit breaker. A lot of research papers have been published related to arc fault of wiring and its detection method. Furthermore, arc fault circuit interrupters have been put into practical use. The objective of the present paper is to understand fault of damaged power supply cord under two selected situations observed in practical use or considered suitable to understand fire ignition. Using two kinds of samples similar to but different from samples prescribed in UL1699 standard, the ignition mechanism of combustibles is discussed based on the results of laboratory experiments. The findings herein underscore the important role of the arc in the ignition of combustibles that are placed on the damaged part of a power supply cord, which is normally followed by a short circuit of broken element conductors or breakage of intact element conductors. Moreover, a possible arc detection feature in the two situations is discussed based on a distorted voltage waveform.

Keywords: arc; AC power supply cord; ignition of combustibles; short circuit; breakage of intact element conductors

1. Introduction Statistics show that, out of the 3972 fire incidents that occurred within the jurisdiction of the Tokyo Fire Department in Japan in 2018, 1205 originated from electrical equipment in domestic and industrial environments [1]. Moreover, the total number of annual electric fire incidents has increased gradually in the past five consecutive years (2014–2018), as depicted in Figure1, eventually reaching 30.3% in 2018 [1]. Table1 points out that one of the major causes of these fires was the wirings of electrical equipment, which include power supply cords and interior wirings. In fact, it has been reported that 38 out of the 448 fire occurrences that broke out from electrical appliances were due to faulty power supply cords, mainly from damaged or deteriorated electrical insulation, disconnection of element conductors by mechanical force, and/or aging [1].

Energies 2020, 13, 681; doi:10.3390/en13030681 www.mdpi.com/journal/energies Energies 2020, 13, 681 2 of 15 Energies 2020, 13, x FOR PEER REVIEW 2 of 15

FigureFigure 1. 1. MajorMajor causes ofof ﬁrefire accidents accidents within within the the jurisdiction jurisdiction of Tokyoof Tokyo Fire Fire Department Department in Japan in Japan from from2014 2014 to 2018 to 2018 [1]. [1].

TableTable 1. 1. OriginOrigin of of fires ﬁres from from el electricalectrical equipment equipment [1]. [1].

OriginOrigin Number Number of ofFires Fires Percentage Percentage (%) (%) ElectricalElectrical appliances appliances 448448 37.2 37.2 (microwaves,(microwaves, rechargeable rechargeable batteries, batteries, etc.) etc.) Wiring Wiring 260 21.6 (power supply cords, interior wirings, etc.) 260 21.6 (power supplyElectrothermal cords, interior apparatus wirings, etc.) 222 18.4 (electricElectrothermal heaters, cooking apparatus heaters, etc.) Wiring accessories 222 18.4 (electric heaters, cooking heaters, etc.) 202 16.8 Wiring(plugs, accessories outlets, etc.) Electric apparatus 202 16.8 68 5.6 (low-voltage(plugs, capacitors, outlets, transformers,etc.) etc.) Electric apparatusOthers 5 0.04 68 5.6 (low-voltage capacitors, transformers, etc.) Others 5 0.04 Some or all of the element conductors in a power supply cord get broken when the cord is trampled by a heavy material, caught in something, or subjected to repetitive bending and stretching. External A lot of research related to the arc fault of wire and its detection method has been carried out force acting on a cord during a large earthquake is also a possible cause [2]. The resultant short circuit earnestly. Standards of AFCI (arc-fault circuit interrupter) or AFDD (arc-fault detection device) of such a damaged power supply cord usually does not trip on a conventional electro-mechanical have been established [5,6]. In studies carried out in accordance with UL1699 standard [7,8], the circuit breaker installed in a distribution panel, because the magnitude and duration of this short-circuit sample is prepared by slicing polyvinylchloride (PVC) across the wire to reveal the conductor, and current are below the operation threshold [3,4]. then, the wire is wrapped with two layers of electrical tape and two layers on fiberglass cloth tape. A lot of research related to the arc fault of wire and its detection method has been carried out A conducting path across the insulation at the slit is created using a high voltage power source. earnestly. Standards of AFCI (arc-fault circuit interrupter) or AFDD (arc-fault detection device) have Parallel arcing is discussed with the sample. In a series arcing sample, the wire of one cable is been established [5,6]. In studies carried out in accordance with UL1699 standard [7,8], the sample is completely broken and there is a space between conductors. It is considered that an arc progresses prepared by slicing polyvinylchloride (PVC) across the wire to reveal the conductor, and then, the wire along carbonized PVC surface or in the space between broken conductors. It has also been reported is wrapped with two layers of electrical tape and two layers on ﬁberglass cloth tape. A conducting path that fire starts from volatile gases produced by decomposition of PVC and the arc continues with across the insulation at the slit is created using a high voltage power source. Parallel arcing is discussed the help of carbon produced by the arc [8]. Furthermore, phenomenon of short circuit in damaged with the sample. In a series arcing sample, the wire of one cable is completely broken and there is a power supply cords or wirings was also studied [9–12], along with the possible methods for its space between conductors. It is considered that an arc progresses along carbonized PVC surface or detection. in the space between broken conductors. It has also been reported that ﬁre starts from volatile gases An arc is characterized by reduced current, voltage drop across the arc, high frequency chatter, produced by decomposition of PVC and the arc continues with the help of carbon produced by the zero-current section in each half cycle of current, steep rise of current, and so on [3]. Several types of arc [8]. Furthermore, phenomenon of short circuit in damaged power supply cords or wirings was also AFCI (Branch/Feeder, outlet circuit, and combination) have been realized [3]. An AFCI seems to studied [9–12], along with the possible methods for its detection. detect frequency component around 100 kHz of current and break the circuit when the component An arc is characterized by reduced current, voltage drop across the arc, high frequency chatter, sustains for more than a few milliseconds [13]. In order to detect a series arc, a band-pass filter was zero-current section in each half cycle of current, steep rise of current, and so on [3]. Several types of proposed with the use of a discrete wavelet transform [14]. A discrete wavelet transform and a AFCI (Branch/Feeder, outlet circuit, and combination) have been realized [3]. An AFCI seems to detect radial basis function neural network were also proposed to identify the occurrence of a series arc in frequency component around 100 kHz of current and break the circuit when the component sustains for indoor low-voltage power lines [15]. In order to identify arc faults from load states, an algorithm is introduced based on a weighted least squares support vector machine [16]. Another power supply

Energies 2020, 13, 681 3 of 15 more than a few milliseconds [13]. In order to detect a series arc, a band-pass filter was proposed with the use of a discrete wavelet transform [14]. A discrete wavelet transform and a radial basis function neural network were also proposed to identify the occurrence of a series arc in indoor low-voltage power lines [15]. In order to identify arc faults from load states, an algorithm is introduced based on a weighted least squares support vector machine [16]. Another power supply interruption method, by converting a short-circuit fault into a line-to-ground fault, was also proposed [17]. Additionally, detecting ultraviolet light, which is emitted by discharge during short circuit, was mentioned as an effective technique [18], whereas another proposed method focused on the distortion of voltage waveform during short circuit [4]. It is worth mentioning that providing a highly reliable fire prevention system requires a better understanding of the physical mechanism of fire occurrences in all prospected scenarios. In line with the arc in damaged power supply cords, two possible physical mechanisms of the ignition of combustibles on the cord are discussed in this paper based on the experimental results using two kinds of samples similar to but different from samples prescribed in UL1699 standard. Essentially, short circuit occurs when a broken element conductor in one cable of a cord comes into contact with an element conductor in another cable, followed by an arc. An arc is also observed when an intact element conductor is melted and disconnected by Joule heating. This paper also provides a possible feature based on a distorted voltage waveform, which can potentially be used to detect an arc in the two situations.

2. Arc at Short Circuits Caused by Contact of Broken Element Conductors

2.1. Sample A commercially available AC power supply cord (125 Vrms, 7 Arms, 0.75 mm2 nominal cross section) was considered for the arc test. Each cable of the cord had 30 bundled copper element conductors (0.18 mm in diameter), with polyvinylchloride (PVC) acting as the electrical insulation. A bending test of the cords was performed with a bending tester (Yasuda Seiki Seisakusyo Ltd., No.225, Nishinomiya, Japan) in accordance with the Japanese Industrial Standard [19]. Conditions of the bending test are summarized in Table2. The test was conducted so that there was no apparent damage on the surface of the PVC. Samples with about 26 broken element conductors in both cables of the cord could be prepared under the conditions. In order to increase the probability of ignition of combustibles, bending test was performed at two locations of a cord, which were separated by a distance of 5 cm.

Table 2. Bending test conditions.

Number of bending times 180 (reciprocation) Bend radius 2.5 mm Bending rate 10 times/min Load 500 g

Appearance of the cord after bending test is shown in Figure2. The white marks on the cord in Figure2 do not indicate damage, but are indication of the location of bending. A microfocus X-ray computed tomography (CT) system (Shimadzu Corporation, SMX-225CT) enabled the nondestructive inspection of a sample. Condition of the element conductors in the sample was observed clearly without removing the PVC. Figure3 shows an example of X-ray CT image. Energies 2020, 13, x FOR PEER REVIEW 3 of 15

interruption method, by converting a short-circuit fault into a line-to-ground fault, was also proposed [17]. Additionally, detecting ultraviolet light, which is emitted by discharge during short circuit, was mentioned as an effective technique [18], whereas another proposed method focused on the distortion of voltage waveform during short circuit [4]. It is worth mentioning that providing a highly reliable fire prevention system requires a better understanding of the physical mechanism of fire occurrences in all prospected scenarios. In line with the arc in damaged power supply cords, two possible physical mechanisms of the ignition of combustibles on the cord are discussed in this paper based on the experimental results using two kinds of samples similar to but different from samples prescribed in UL1699 standard. Essentially, short circuit occurs when a broken element conductor in one cable of a cord comes into contact with an element conductor in another cable, followed by an arc. An arc is also observed when an intact element conductor is melted and disconnected by Joule heating. This paper also provides a possible feature based on a distorted voltage waveform, which can potentially be used to detect an arc in the two situations.

2. Arc at Short Circuits Caused by Contact of Broken Element Conductors

2.1. Sample

A commercially available AC power supply cord (125 Vrms, 7 Arms, 0.75 mm2 nominal cross section) was considered for the arc test. Each cable of the cord had 30 bundled copper element conductors (0.18 mm in diameter), with polyvinylchloride (PVC) acting as the electrical insulation. A bending test of the cords was performed with a bending tester (Yasuda Seiki Seisakusyo Ltd., No.225, Nishinomiya, Japan) in accordance with the Japanese Industrial Standard [19]. Conditions of the bending test are summarized in Table 2. The test was conducted so that there was no apparent damage on the surface of the PVC. Samples with about 26 broken element conductors in both cables of the cord could be prepared under the conditions. In order to increase the probability of ignition of combustibles, bending test was performed at two locations of a cord, which were separated by a distance of 5 cm. Appearance of the cord after bending test is shown in Figure 2. The white marks on the cord in Figure 2 do not indicate damage, but are indication of the location of bending. A microfocus X-ray computed tomography (CT) system (Shimadzu Corporation, SMX-225CT) enabled the nondestructive inspection of a sample. Condition of the element conductors in the sample was observed clearly without removing the PVC. Figure 3 shows an example of X-ray CT image. Herein, “partially broken” is used to describe the status in which some of the element conductors are broken and at least one element conductor remains intact.

Table 2. Bending test conditions.

Number of bending times 180 (reciprocation) Bend radius 2.5 mm Bending rate 10 times/min Energies 2020, 13, 681 Load 500 g 4 of 15

Energies 2020, 13, x FOR PEER REVIEWFigure 2. A typical power supply cord after bending test. 4 of 15

(a) (b)

FigureFigure 3. 3.ExampleExample of ofX-ray X-ray computed computed tomography tomography image. image. (a)( aoverall) overall image image (b) ( bmagnified) magniﬁed image image of of circledcircled part part in in(a). (a ).

TheHerein, present “partially sample broken”is similar is to used a sample to describe descr theibed status in UL1699 in which standard someof [5], the which element is used conductors for a parallelare broken arc experiment. and at least oneThe element latter conductoris prepared remains by slicing intact. PVC across the wire to reveal the conductorThe presentand then sample the wire is similar is wrapped to a sample with described two layers in UL1699of electrical standard tape [ 5and], which two islayers used forof a fiberglassparallel arccloth experiment. tape. Conducting The latter path is preparedacross the by insulation slicing PVC at the across slit is the created wire tousing reveal a high theconductor voltage powerand thensource the [7]. wire Conversely, is wrapped PVC with is not two sliced layers in of the electrical present tapesample, and which two layers is probably of fiberglass close to cloth a practicaltape. Conducting situation. path across the insulation at the slit is created using a high voltage power source [7]. Conversely, PVC is not sliced in the present sample, which is probably close to a practical situation. 2.2. Experimental Procedures 2.2. Experimental Procedures The experimental circuit is shown in Figure 4. Here, an AC voltage of 100 Vrms and 60 Hz was suppliedThe to experimental the sample from circuit a isdistribution shown in Figureboard4 in. Here, the laboratory. an AC voltage The ofresistance 100 Vrms of and the 60 internal Hz was wiringssupplied was to simula the sampleted with from two a 0.4 distribution Ω resistors. board A 45 incm the long laboratory. sample cord The was resistance laid straight of the and internal flat, withwirings the load was being simulated eight with incandescent two 0.4 Ω lampsresistors. (160 A0 45W cm in longtotal). sample Combustibles cord was cut laid from straight a cushion and flat, werewith placed the load on the being sample eight to incandescent cover the two lamps partia (1600lly broken W in total).parts. CombustiblesThe principal component cut from a cushionof the combustibleswere placed was on thedry samplecotton. to cover the two partially broken parts. The principal component of the combustiblesThe voltage was and dry current cotton. waveforms were recorded simultaneously using a digital oscilloscope (TektronixThe voltageTDS3034B). and currentSubsequently, waveforms these were voltage recorded and current simultaneously data were using applied a digital for calculating oscilloscope the(Tektronix resistances TDS3034B). of the cables. Subsequently, An infrared these thermometer voltage and (Keysight current dataU5855A) were and applied a video for camera calculating were the alsoresistances used. Based of the on cables.the preliminary An infrared experiments, thermometer the (Keysighttemperature U5855A) on the and PVC a videosurface camera increased were just also afterused. the Based commencement on the preliminary of current experiments, flow and thetended temperature to saturate on theafter PVC 10 min. surface About increased 10 min just was after enoughthe commencement for the sample of current to be flowcooled and tendeddown to saturateambient after temperature 10 min. About [12]. 10 Considering min was enough these for conditions,the sample a to60 be min cooled cycle down was toadopted ambient in temperature the experiment, [12]. Consideringin which a these13.5 Arms conditions, current a 60was min suppliedcycle was to adoptedthe sample in the for experiment, 45 min, followed in which by a 13.5a pause Arms for current 15 min. was The supplied cycleto was the samplerepeated for intermittently.45 min, followed by a pause for 15 min. The cycle was repeated intermittently.

Figure 4. Experimental circuit for samples with the partially broken element conductors.

2.3. Results and Discussion

Energies 2020, 13, x FOR PEER REVIEW 4 of 15

(a) (b)

Figure 3. Example of X-ray computed tomography image. (a) overall image (b) magnified image of circled part in (a).

The present sample is similar to a sample described in UL1699 standard [5], which is used for a parallel arc experiment. The latter is prepared by slicing PVC across the wire to reveal the conductor and then the wire is wrapped with two layers of electrical tape and two layers of fiberglass cloth tape. Conducting path across the insulation at the slit is created using a high voltage power source [7]. Conversely, PVC is not sliced in the present sample, which is probably close to a practical situation.

2.2. Experimental Procedures The experimental circuit is shown in Figure 4. Here, an AC voltage of 100 Vrms and 60 Hz was supplied to the sample from a distribution board in the laboratory. The resistance of the internal wirings was simulated with two 0.4 Ω resistors. A 45 cm long sample cord was laid straight and flat, with the load being eight incandescent lamps (1600 W in total). Combustibles cut from a cushion were placed on the sample to cover the two partially broken parts. The principal component of the combustibles was dry cotton. The voltage and current waveforms were recorded simultaneously using a digital oscilloscope (Tektronix TDS3034B). Subsequently, these voltage and current data were applied for calculating the resistances of the cables. An infrared thermometer (Keysight U5855A) and a video camera were also used. Based on the preliminary experiments, the temperature on the PVC surface increased just after the commencement of current flow and tended to saturate after 10 min. About 10 min was enough for the sample to be cooled down to ambient temperature [12]. Considering these conditions, a 60 min cycle was adopted in the experiment, in which a 13.5 Arms current was Energiessupplied2020 ,to13 ,the 681 sample for 45 min, followed by a pause for 15 min. The cycle was repeated5 of 15 intermittently.

FigureFigure 4.4. Experimental circuit for samples withwith thethe partiallypartially brokenbroken elementelement conductors.conductors. Energies 2020, 13, x FOR PEER REVIEW 5 of 15 2.3.2.3. Results and Discussion Figure 5 illustrates the temperature variation on the PVC surface with respect to the number of Figure5 illustrates the temperature variation on the PVC surface with respect to the number of cycles. Here, temperature means the highest value reached in a cycle. At the beginning, the temperaturecycles. Here, increased temperature gradually means the with highest the number value reached of cycles. in a Then, cycle. a At sudden the beginning, increase the was temperature distinctly observedincreased graduallyduring the with 10th the cycle, number when of cycles. the Then,temperature a sudden rose increase to approximately was distinctly observed180 °C, duringbefore becomingthe 10th cycle, almost when constant the temperature afterwards. rose A remarkable to approximately variation 180 in◦C, the before resistance becoming of each almost cable constant could alsoafterwards. be confirmed A remarkable at the 10th variation cycle; in it theis worth resistance noting of each that cablesuch couldvariation also of be resistance conﬁrmed was at the slight 10th beforecycle; itand is worthafter the noting 10th thatcycle. such Similar variation changes of resistancein the temperature was slight and before resistance and after of cables, the 10th such cycle. as theSimilar ones changesdepicted in in the Figure temperature 5, were observed and resistance in other of samples. cables, such as the ones depicted in Figure5, were observed in other samples.

FigureFigure 5.5. TemperatureTemperature on on the the polyvinylchloride polyvinylchloride (PVC) (PVC) near thenear partially the partially broken partbroken and thepart resistances and the resistancesof the cables ofwith the cables the partially with the broken partially element broken conductors. element conductors.

FigureFigure 66 showsshows temperaturetemperature distributiondistribution aroundaround thethe partiallypartially brokenbroken partpart measuredmeasured withwith anan infraredinfrared thermometer. NoteNote thatthat thethe steepsteep rise rise in in temperature temperature of of the the power power supply supply cord cord is is observed observed in ina limiteda limited region, region, as shown as shown in Figure in6 a.Figure The temperature 6a. The temperature on the outside on surfacethe outside of the combustiblessurface of the is combustiblesalmost the same is almost as the ambientthe same temperature, as the ambient as shown temperature, in Figure as6b. shown Moreover, in Figure Figure 6b.6c showsMoreover, the Figuretemperature 6c shows distribution the temperature when the combustibles distribution are wh turneden the over. combustibles It is conﬁrmed are thatturned the temperatureover. It is confirmedof the power that supply the temperature cord and the of surface the power of the suppl combustiblesy cord and in contact the surface with itof reaches the combustibles approximately in contact200 ◦C. with it reaches approximately 200 °C. FigureFigure 77 showsshows X-rayX-ray computedcomputed tomographytomography images,images, whichwhich clearlyclearly showshow thethe conditioncondition ofof thethe elementelement conductors atat aa partially partially broken broken part part of of the the sample sample before before the the start start of the of experimentthe experiment and afterand afterthe 6th the cycle 6th hascycle been has completed. been completed. The broken The element broken conductors element conductors were tightly were bundled tightly together bundled and togetherassumed and contact assumed with othercontact element with other conductors element prior conductors to the onset prior of theto the experiment. onset of the On experiment. the contrary, Onthe the element contrary, conductors the element were conductors seemingly were loosely seemingly bundled loosely after being bundled subjected after being to a temperature subjected to ofa temperature of approximately 150 °C for six cycles. A large space was also found between the broken element conductors. Considering that the glass transition temperature of PVC is within 70– 80 °C [20], these changes in the condition of the element conductors have probably been formed by thermal expansion and shrinkage of PVC during the cycles. Although different samples were used in Figures 5 and 7, it is suggested that the sudden increase in temperature and resistance, as illustrated in Figure 5, may be explained by such a change in the condition of the element conductors.

Energies 2020, 13, 681 6 of 15

approximately 150 ◦C for six cycles. A large space was also found between the broken element conductors. Considering that the glass transition temperature of PVC is within 70–80 ◦C[20], these changes in the condition of the element conductors have probably been formed by thermal expansion and shrinkage of PVC during the cycles. Although diﬀerent samples were used in Figures5 and7, it is suggested that the sudden increase in temperature and resistance, as illustrated in Figure5, may be EnergiesexplainedEnergies 2020 2020, by13, ,13 x such ,FOR x FOR PEER a changePEER REVIEW REVIEW in the condition of the element conductors. 6 of6 of15 15

(a)( a) (b() b)

(c)( c)

FigureFigureFigure 6. 6. Temperature6.Temperature Temperature distribution distribution distribution around around around the the the partially partially partially broken broken broken part part part (Voltage (Voltage (Voltage 100 100 100 Vrms, Vrms, Vrms, Current Current Current 13.5 13.5 13.5 Arms).Arms).Arms). (a ( )a( a)External) External External appearance appearance appearance (front (front (front view); view); view); (b ( b)( b)External) External External appearance appearance appearance (oblique (oblique (oblique view); view); view); (c ( )c( )cSurface) Surface Surface of of of combustiblescombustiblescombustibles and and and power power power supply supply supply cord cord cord (combustibles (combustibles (combustibles are are are removed removed removed and and and turned turned turned over). over). over).

(a)( aBefore) Before the the experimen experiment t (b()b After) After six six cycles cycles have have been been completed completed AppearanceAppearance

ImageImage of ofneutral neutral cable cable

ImageImage of ofline line cable cable

Figure 7. Appearance of partially broken power supply cords and images of element conductors in cords obtained with the microfocus X-ray CT system. FigureFigure 7. 7.Appearance Appearance of ofpartially partially brok brokenen power power supply supply cords cords and and images images of ofelement element conductors conductors in in cordscords obtained obtained with with the the microfocus microfocus X-ray X-ray CT CT system. system.

ForFor the the sample sample shown shown in inFigure Figure 5, 5,the the arc arc oc occurredcurred just just after after the the commencement commencement of ofcurrent current flowflow in inthe the 31st 31st cycle; cycle; thus, thus, ignition ignition of ofthe the comb combustiblesustibles in inFigure Figure 8 was8 was observed. observed. At At the the partially partially brokenbroken part part of ofthe the sample, sample, the the current current concentrates concentrates on on the the remaining remaining intact intact element element conductors conductors and and causescauses the the generation generation of ofa alarge large amount amount of ofJo uleJoule heat. heat. Accordingly, Accordingly, the the PVC PVC is issoftened softened and and expanded.expanded. Because Because the the PVC PVC contracts contracts in inthe the cooling cooling process process when when the the current current flow flow pauses, pauses, a brokena broken elementelement conductor conductor in inone one cable cable of ofthe the cord cord may may pierce pierce and and break break through through the the PVC PVC between between cables cables

Energies 2020, 13, 681 7 of 15

For the sample shown in Figure5, the arc occurred just after the commencement of current ﬂow in the 31st cycle; thus, ignition of the combustibles in Figure8 was observed. At the partially broken part of the sample, the current concentrates on the remaining intact element conductors and causes the generation of a large amount of Joule heat. Accordingly, the PVC is softened and expanded. Because the PVC contracts in the cooling process when the current ﬂow pauses, a broken element Energies 2020, 13, x FOR PEER REVIEW 7 of 15 conductor in one cable of the cord may pierce and break through the PVC between cables by shrinkage stress. When the element conductor comes into contact with that in the other cable, a short circuit by shrinkage stress. When the element conductor comes into contact with that in the other cable, a occurs by bridging conductors of cables. Thereby, an arc may be generated when current is supplied short circuit occurs by bridging conductors of cables. Thereby, an arc may be generated when again in the next cycle. This hypothesis is probably supported by a small hole found sometimes in current is supplied again in the next cycle. This hypothesis is probably supported by a small hole thermally deformed PVC. found sometimes in thermally deformed PVC.

Figure 8. Ignition of combustibles. Figure 8. Ignition of combustibles.

TheThe resultsresults of of the the eight eight samples samples are summarizedare summarize in Tabled in 3Table. Here, 3. the Here, word the “temperature” word “temperature” indicates theindicates highest the temperature highest temperature recorded by recorded the end of by the th experimente end of the of aexperiment sample, whereas of a sample, “commencement whereas time”“commencement denotes the time” time whendenotes the the short time circuit when occurred the short after circuit current occurred has been after supplied current in thehas cycle.been Insupplied most samples, in the cycle. short In circuit most occurred samples, immediately short circuit or occurred shortly after immediately the cycle currentor shortly ﬂow after commenced, the cycle whichcurrent suggests flow commenced, that the broken which elementsuggests conductors that the broken of both element cables conductors contact with of both each cables other duringcontact thewith cooling each other process during in the the previous cooling process cycle, as in mentioned the previous above. cycle, Such as mentioned a short circuit above. continued Such a short for a rathercircuit longcontinued time or repeatedfor a rather intermittently long time until or allrepeated of the element intermittently conductors until were all broken. of the It iselement noted thatconductors an arc during were broken. the short It circuitis note didd that not an always arc during lead tothe the short ignition circuit of combustibles.did not always The lead thermal to the circuitignition breaker of combustibles. (20 Arms) that The was thermal installed circuit in thebreake circuitr (20 did Arms) not trip that in was these installed experiments, in the because circuit did the short-circuitnot trip in these current experiments, and its duration, because as the shown short-circ in Tableuit3 ,current were below and its the duration, threshold as values shown of operation.in Table 3, were below the threshold values of operation. Table 3. Experimental results for the partially broken samples. Table 3. Experimental results for the partially broken samples. Characteristics of the Short Circuit Sample No. Temperature (◦C) Characteristics of the Short Circuit Temperature No. of Cycles Commencement Time Current (Arms) Duration (ms) Sample No. Commencement Current Duration 1 175.2(°C) No. 31of Cycles 122ms 80.6 1,070 2 212.4 25 67.1msTime 71.9(Arms) 240(ms) 13 225.0175.2 3831 30.0ms 122ms 86.180.6 5601,070 4 223.0 5 10min 87.6 850 25 224.0212.4 1525 Immediately 67.1ms 51.271.9 80240 36 231.0225.0 538 Immediately 30.0ms 30.086.1 21560 47 213.0223.0 13 5 25.3ms 10min 93.487.6 253850 8 186.6 23 Immediately 44.5 83 5 224.0 15 Immediately 51.2 80 6 231.0 5 Immediately 30.0 21 It7 has been reported213.0 that ﬁre starts13 from volatile gases 25.3ms produced by decomposition93.4 of PVC253 and the arc8 continues with186.6 the help of the23 carbon producedImmediately by the arc [8]. In the44.5 present experiment,83 short circuit mostly occurs, and the arc generates immediately or shortly after the commencement of It has been reported that fire starts from volatile gases produced by decomposition of PVC and current supply. It is considered temperature of PVC is not so high to produce volatile gases. Thus, the arc continues with the help of the carbon produced by the arc [8]. In the present experiment, the following scenario may be proposed. short circuit mostly occurs, and the arc generates immediately or shortly after the commencement (1)of currentMechanical supply. forces It is breakconsidered some oftemperature the element of conductors, PVC is not thus so high allowing to produce the formation volatile of gases. a partially Thus, the followingbroken part scenario in the may power be supplyproposed. cord. (1) Mechanical forces break some of the element conductors, thus allowing the formation of a partially broken part in the power supply cord. (2) At the partially broken part, the current runs through the remaining intact element conductors, which results in a rise in temperature of the PVC due to a large amount of Joule heat. (3) The softened PVC releases a tightening force applied to the element conductors. (4) The broken element conductor in one cable of the cord pierces and breaks through the PVC between cables and comes into contact with other element conductors in the other cable,

Energies 2020, 13, 681 8 of 15

(2) At the partially broken part, the current runs through the remaining intact element conductors, which results in a rise in temperature of the PVC due to a large amount of Joule heat. (3)EnergiesThe 2020 softened, 13, x FOR PVC PEER releases REVIEW a tightening force applied to the element conductors. 8 of 15 (4) The broken element conductor in one cable of the cord pierces and breaks through the PVC between probably by shrinkage stress acting on PVC during the cooling process when the current cables and comes into contact with other element conductors in the other cable, probably by supply is paused. shrinkage stress acting on PVC during the cooling process when the current supply is paused. (5) Since the element conductor bridges conductors of two cables by the end of a cooling process, a (5) shortSince circuit the element occurs conductor immediately bridges or shortly conductors when of current two cables flows by again the end in the of a next cooling cycle, process, and an a arcshort generates. circuit occurs immediately or shortly when current ﬂows again in the next cycle, and an (6) Byarc coming generates. into contact with an arc or a piece of melted element conductor, a minute portion of (6) theBy comingcombustible into contact occasionally with an reaches arc or a its piece flash of po meltedint temperature, element conductor, which results a minute in portion probabilistic of the ignition.combustible occasionally reaches its ﬂash point temperature, which results in probabilistic ignition. In realreal life life situations, situations, these these phenomena phenomena are possibly are possibly observed observed in a power in a supply power cord supply of household cord of applianceshousehold appliances and in an extensionand in an cordextension when cord they when are trampled they are trampled by heavy by material heavy suchmaterial asa such chest as of a drawers,chest of drawers, caught incaught something, in something, or subjected or subjec to repetitiveted to repetitive bending bending and stretching. and stretching. There There may may be a possibilitybe a possibility to observe to observe these phenomenathese phenomena in cords in subjected cords subjected to frequent to frequent heat cycle, heat for example,cycle, for causedexample, by repetitioncaused by ofrepetition supplying of andsupplying pausing and current. pausing More current. investigation More investigation will be necessary will be to necessary validate the to proposedvalidate the scenario. proposed scenario.

3. Arc Caused by the Breakage of Intact Element Conductor

3.1. Sample The tested sample was a commercially availa availableble AC power supply supply cord cord (300 (300 Vrms, Vrms, 12 Arms, 2 1.25 mmmm2 nominal cross section). Each cable of the cordcord had 50 bundled copper element conductors (0.18 mmmm inin diameter),diameter), withwith PVCPVC actingacting asas thethe electricalelectrical insulation.insulation. Two cablescables ofof the the sample sample cord cord were were separated separated from from each each other. other. After After partially partially removing removing the PVC the ofPVC one of of one the of cables, the cables, 49 element 49 element conductors conductors were broken were artiﬁciallybroken artificially and also and removed. also removed. Thus, only Thus, one elementonly one conductor element conductor remained remained intact, as shownintact, inas Figureshown9 .in No Figure treatment 9. No was treatment given in was the given other cable.in the Theother sample cable. wasThe 10sample cm long, was and10 cm its long, partially and brokenits partially part wasbroken 1 cm part long. was 1 cm long.

Figure 9. Appearance of a sample with one intact element conductor.

This sample is similar to a sample used for seriesseries arcing experimentexperiment [[7],7], where conductorconductor is completely brokenbroken andand there there is is a spacea space between between conductors. conductors. It is It considered is considered that arcthat progresses arc progresses along PVCalong surface PVC surface or in the or spacein the betweenspace between broken broken conductors. conductors. On the On contrary, the contrary, in the present in the sample,present elementsample, conductorselement conductors are not completely are not completely broken and br oneoken remains and intact.one remains An arc intact. generates An whenarc generates an intact conductorwhen an intact element conductor is melted element by Joule is melted heating. by Joule heating.

3.2. Experimental Procedures The experimental circuit is shown in FigureFigure 1010.. The The sample sample was was connected connected to to a a load of 10 incandescent lampslamps (1000 (1,000 W inW total), in total), with 100with Vrms 100 (60 Vrms Hz) and(60 9.7Hz) Arms and being 9.7 Arms supplied being continuously supplied tocontinuously the sample fromto the a sample wall outlet from in a the wall laboratory. outlet in Asthe the laboratory. ignition ofAs the the combustibles ignition of the was combustibles investigated bywas arc investigated discharge, theyby arc were discharge, placed on they the samplewere pl toaced cover on thethe intact sample element to cover conductor, the intact as shown element in conductor, as shown in Figure 11. The principal component of the combustibles was dry cotton cut from a cushion: the same material used in the experiments in Chapter II. Voltage and current waveforms were recorded with 1 μs sampling time using a digital oscilloscope (Agilent 54832B) until the intact element conductor was melted by Joule heating. A high-speed camera (Photron FASTCAM SA1.1) was used to record the phenomenon.

Energies 2020, 13, 681 9 of 15

Figure 11. The principal component of the combustibles was dry cotton cut from a cushion: the same material used in the experiments in Chapter II. EnergiesEnergies 20202020,, 1313,, xx FORFOR PEERPEER REVIEWREVIEW 99 ofof 1515

Energies 2020, 13, x FOR PEER REVIEW 9 of 15

Figure 10. Experimental circuit for sample with one intact element conductor. FigureFigure 10.10. ExperimentalExperimental circuitcircuit forfor samplesample withwith oneone intactintact elementelement conductor.conductor.

Figure 10. Experimental circuit for sample with one intact element conductor.

FigureFigureFigure 11. 11.11. CombustiblesCombustibles placedplaced onon thethe samplesample cord.cord.cord.

3.3.3.3. ResultsResultsVoltage andand and DiscussionDiscussion current waveforms were recorded with 1 µs sampling time using a digital oscilloscope (Agilent 54832B) until the intact element conductor was melted by Joule heating. A high-speed camera FigureFigure 1212 showsshows aa photographphotograph atat thethe timetime ofof ararcc generationgeneration inin thethe casecase ofof nono combustibles.combustibles. (Photron FASTCAM SA1.1) was used to record the phenomenon. ContinuousContinuous currentcurrent flowingflowingFigure inin thethe 11. intact intactCombustibles elementelement placed conductorconductor on the resultedresultedsample cord. inin aa riserise inin itsits temperaturetemperature byby Joule heating. Finally, the element conductor melted and arc occurred. The arc ceased within a few 3.3.Joule Results heating. and Finally, Discussion the element conductor melted and arc occurred. The arc ceased within a few 3.3. Results and Discussion millisecondsmilliseconds withwith thethe breakagebreakage ofof thethe elementelement condconductor.uctor. DispersionDispersion ofof high-temperaturehigh-temperature piecespieces ofof Figure 12 shows a photograph at the time of arc generation in the case of no combustibles. thethe meltedmeltedFigure elementelement12 shows conductorconductor a photograph isis confirmedconfirmed at the inintime FigureFigure of ar 12.12.c generation in the case of no combustibles. Continuous current ﬂowing in the intact element conductor resulted in a rise in its temperature by Continuous current flowing in the intact element conductor resulted in a rise in its temperature by Joule heating. Finally, the element conductor melted and arc occurred. The arc ceased within a few Joule heating. Finally, the element conductor melted and arc occurred. The arc ceased within a few milliseconds with the breakage of the element conductor. Dispersion of high-temperature pieces of the milliseconds with the breakage of the element conductor. Dispersion of high-temperature pieces of melted element conductor is conﬁrmed in Figure 12. the melted element conductor is confirmed in Figure 12.

FigureFigure 12.12. ArcArc observedobserved atat thethe timetime ofof breakagebreakage ofof anan elementelement conductor.conductor.

WhenWhen therethere werewere combustibles,combustibles, ththeireir ignition,ignition, asas illustratedillustrated inin FigureFigure 13,13, couldcould bebe observedobserved occasionally.occasionally. However,However, notenote thatthat thethe probabilityprobability ofof ignitionignition ofof thethe combustiblescombustibles waswas muchmuch lowerlower than that described in the experiment in Chapter II, which could be attributed to the weak and than that describedFigure in 12.the Arc experiment observed at in the Chapte time ofr breakageII, which of couldan element be attributed conductor. to the weak and Figure 12. Arc observed at the time of breakage of an element conductor. short-durationshort-duration arcarc inin thethe presentpresent experiment.experiment. Nevertheless,Nevertheless, thethe ignitionignition probabilityprobability mightmight increaseincrease with the increase in the number of intact element conductors. The thermal circuit breaker (20 Arms) withWhen the increase there werein the combustibles,combustibles, number of intact theirtheir elemen ignition,t conductors. as illustrated The thermal in Figure circuit 1313,, could breaker bebe (20 observedobserved Arms) installed in the distribution panel did not trip. occasionally.installed in the However,However, distribution note note thatpanel that the didthe probability notprobability trip. of ignitionof ignition of theof the combustibles combustibles was was much much lower lower than thatthan described that described in the experimentin the experiment in Chapter in II,Chapte whichr couldII, which be attributed could be to attributed the weak and to short-durationthe weak and arcshort-duration in the present arc experiment. in the present Nevertheless, experiment. the Nevertheless, ignition probability the ignition might probability increase with might the increase inwith the the number increase of intactin the elementnumber of conductors. intact elemen Thet conductors. thermal circuit The breaker thermal (20 circuit Arms) breaker installed (20 Arms) in the distributioninstalled in the panel distribution did not trip. panel did not trip.

FigureFigure 13.13. IgnitionIgnition ofof combustibles.combustibles.

AA possiblepossible ignitionignition mechanismmechanism ofof combustiblescombustibles causedcaused byby arcarc atat thethe timetime ofof breakagebreakage ofof anan intactintact elementelement conductorconductor cancan bebe describeddescribed asas follows:follows: (1)(1) An An externalexternal mechanicalmechanical forceforce Figure breaksbreaks 13. mostmost Ignition ofof thethe of elementcombustibles.element conductorsconductors ofof aa powerpower supplysupply cord.cord.

A possible ignition mechanism of combustibles caused by arc at the time of breakage of an intact element conductor can be described as follows: (1) An external mechanical force breaks most of the element conductors of a power supply cord.

Energies 2020, 13, x FOR PEER REVIEW 9 of 15

Figure 10. Experimental circuit for sample with one intact element conductor.

Figure 11. Combustibles placed on the sample cord.

3.3. Results and Discussion Figure 12 shows a photograph at the time of arc generation in the case of no combustibles. Continuous current flowing in the intact element conductor resulted in a rise in its temperature by Joule heating. Finally, the element conductor melted and arc occurred. The arc ceased within a few milliseconds with the breakage of the element conductor. Dispersion of high-temperature pieces of the melted element conductor is confirmed in Figure 12.

Figure 12. Arc observed at the time of breakage of an element conductor.

When there were combustibles, their ignition, as illustrated in Figure 13, could be observed occasionally. However, note that the probability of ignition of the combustibles was much lower than that described in the experiment in Chapter II, which could be attributed to the weak and short-duration arc in the present experiment. Nevertheless, the ignition probability might increase Energieswith the2020 increase, 13, 681 in the number of intact element conductors. The thermal circuit breaker (20 Arms)10 of 15 installed in the distribution panel did not trip.

FigureFigure 13.13. Ignition of combustibles.

A possiblepossible ignitionignition mechanism mechanism of of combustibles combustibles caused caused by arcby atarc the at timethe time of breakage of breakage of an intactof an elementintact element conductor conductor can be can described be described as follows: as follows:

(1)(1) An external mechanical force breaks most of the element conductors of a power supply cord.cord. (2) Current runs through the intact element conductors and its temperature rises because of Joule heating. (3) The intact element conductors melt and arc generates. (4) The temperature of a minute portion of the combustibles occasionally reaches the ﬂash point through contact with arc or a piece of the melted element conductor, which results in the probabilistic ignition of combustibles.

In a real world, this type of damage might be observed when a power supply cord is deteriorated seriously, for example, under thermally and/or chemically severe environment. It is well known that oxidation of insulating materials will result in reduction of their mechanical and electrical properties and that high temperature will generally accelerate degradation of insulating materials.

4. Possible Arc Detection Feature A possible feature will be discussed here, which can potentially be used to detect an arc in the two samples described above. In our previous paper [21], a feature was proposed for detecting an arc at the time of breakage of an intact element conductor under a condition without combustibles. The distortion of a recorded voltage waveform from a sinusoidal voltage was focused on as the characteristic of the arc. Subsequently, the voltage waveform recorded just before arc generation was used as the “checking voltage” for comparison. The checking voltage is the voltage in the absence of arc. It is obtained by the following procedures; (1) the arc is detected by the proposed criteria described below, (2) a half-cycle voltage waveform at arc generation is extracted, (3) a half-cycle voltage waveform of the same polarity with that obtained in (2) just before arc generation is called form recorded data, which is named checking voltage; (4) the two voltage waveforms are superimposed so that the zero cross points coincide with each other. The procedures are illustrated in Figure 14. In the practical case, the checking voltage can be obtained by continuous recording of voltage at an outlet or at a distribution panel. The “deviation time” was also introduced for judging the arc generation [21]. It was deﬁned as the sum of the durations when the instantaneous experimental voltage exceeded 6 V of the instantaneous ± checking voltage. Arc generation is diagnosed when the following two criteria are satisﬁed:

(1) The deviation time is 70 µs or longer in a half cycle. (2) Current becomes zero within the next half cycle.

The ﬁrst criterion is derived from the results of a series of experiments in the laboratory, whereas the second criterion is necessary to prevent incorrect judgment due to transient phenomena, which may cause the distortion of voltage waveforms, for example, when a household appliance is turned on or when its operation condition is changed. However, in these cases, current continues to ﬂow for more than one cycle. Values of 6 V and 70 µs, and a half cycle in the criteria are determined by trial and ± error and their optimization will be necessary through further experiments. Energies 2020, 13, 681 11 of 15

The criteria will be applied to detection of two kinds of arc described in Chapters II and III. To begin with, the case of arc at a short circuit caused by contact of broken element conductors is considered. Figure 15 shows waveforms recorded under a condition similar to that obtained Figure8. The experimental voltage waveform showed a line voltage between line and neutral, which was distorted from the sinusoidal during arc. At arc extinction, the current ceased and voltage got restored to the sinusoidal. The waveform during the arc (surrounded by a green ellipsoid in Figure 15) is shown in a magnified view in Figure 16. Note that the amplitude of the current exceeded 100 A for one period and that the distortion of the voltage waveform is clearly confirmed. Figure 17 is a magnified view of the waveforms that were surrounded by a yellow ellipsoid in Figure 16, which is used to calculate deviation time. Here, the experimental voltage is drawn in red, while the checking voltage and the range of checking voltage 6 V are shown in green and light green, respectively. Moreover, the arrows ± Energiesin Figure 2020 17, 13 indicate, x FOR PEER the REVIEW durations when the experimental voltage has exceeded the checking voltage11 of 15 6 V. The deviation time defined by the sum of the durations is 4300 µs, which is longer than 70 µs. ± This case satisfied both of the criteria above, which implies that detection of the arc is possible.

Figure 14. Procedures to obtain checking voltage. Figure 14. Procedures to obtain checking voltage.

The criteria will be applied to detection of two kinds of arc described in Chapters II and III. To begin with, the case of arc at a short circuit caused by contact of broken element conductors is considered. Figure 15 shows waveforms recorded under a condition similar to that obtained Figure 8. The experimental voltage waveform showed a line voltage between line and neutral, which was distorted from the sinusoidal during arc. At arc extinction, the current ceased and voltage got restored to the sinusoidal. The waveform during the arc (surrounded by a green ellipsoid in Figure 15) is shown in a magnified view in Figure 16. Note that the amplitude of the current exceeded 100 A for one period and that the distortion of the voltage waveform is clearly confirmed. Figure 17 is a magnified view of the waveforms that were surrounded by a yellow ellipsoid in Figure 16, which is used to calculate deviation time. Here, the experimental voltage is drawn in red, while the checking voltage and the range of checking voltage ±6 V are shown in green and light green, respectively. Moreover, the arrows in Figure 17 indicate the durations when the experimental voltage has exceeded the checking voltage ±6 V. The deviation time defined by the sum of the durations is 4300 μs, which is longer than 70 μs. This case satisfied both of the criteria above, which implies that detection of the arc is possible.

Energies 2020, 13, x FOR PEER REVIEW 12 of 15

Energies 2020, 13, x FOR PEER REVIEW 12 of 15 Energies 2020, 13, 681 12 of 15

Figure 15. Sample voltage and current waveforms at a short circuit.

FigureFigure 15.15. Sample voltage and current waveformswaveforms atat aa shortshort circuit.circuit.

Figure 16. Magnified view of the part surrounded by a green ellipsoid in Figure 15. Figure 16. Magniﬁed view of the part surrounded by a green ellipsoid in Figure 15. Figure 16. Magnified view of the part surrounded by a green ellipsoid in Figure 15.

Figure 16. Magnified view of the part surrounded by a green ellipsoid in Figure 15.

Deviation Time

Figure 17. Magnified view of the part surrounded by a yellow ellipsoid in Figure 16. Figure 17. Magnified view of the part surrounded by a yellow ellipsoid in Figure 16. The next discussion will underscore the case of arc at the breakage of the intact element conductor. Figure 17. Magnified view of the part surrounded by a yellow ellipsoid in Figure 16. FigureThe 18 nextshows discussion the waveforms will underscore recorded underthe case a condition of arc at similar the breakage to that inof Figure the intact 13. Basedelement on conductor. Figure 18 shows the waveforms recorded under a condition similar to that in Figure 13. the figure,The next theFigure discussion arc would17. Magnified resultwill underscore view in a distortionof the partthe surroundedincase both of waveforms arc by ata yellowthe withbreakage ellipsoid a smaller inof Figure the degree intact 16. than element in the Based on the figure, the arc would result in a distortion in both waveforms with a smaller degree conductor.first case, because Figure 18 only shows one elementthe waveforms conductor recorded is broken under in thea condition present case. similar The to current that in ceasedFigure and13. than in the first case, because only one element conductor is broken in the present case. The current Basedthe voltageThe on thenext waveform figure, discussion the got arc restoredwill would underscore result to the sinusoidalin thea distortion case atof arc arcin extinction, bothat the waveforms breakage within with aof cycle the a smaller afterintact the elementdegree onset ceased and the voltage waveform got restored to the sinusoidal at arc extinction, within a cycle after thanofconductor. voltage in the distortion.first Figure case, 18 because shows Figure the 19only showswaveforms one element a magnified recorded conductor view under ofis thebrokena condition waveforms in the similar present during to case.that arc (surroundedinThe Figure current 13. ceasedbyBased a green andon the ellipsoidthe figure, voltage in the Figurewaveform arc would18), got in whichresult restored in a distortiona to distortion the sinusoidal of thein both voltage at arcwaveforms waveformextinction, with iswithin clearly a smaller a cycle confirmed. degree after

Thethan deviation in the first time case, could because be obtained only one from element Figure conductor 20, which is broken is a magniﬁed in the present view of case. the waveformsThe current surroundedceased and the by voltage a yellow waveform ellipsoid ingot Figure restored 19. to The the experimental sinusoidal at voltagearc extinction, is drawn within in red, a cycle while after the

Energies 2020, 13, x FOR PEER REVIEW 13 of 15

Energies 2020, 13, x FOR PEER REVIEW 13 of 15 the onset of voltage distortion. Figure 19 shows a magnified view of the waveforms during arc (surrounded by a green ellipsoid in Figure 18), in which a distortion of the voltage waveform is Energiesthe onset 2020 ,of 13 , voltagex FOR PEER distortion. REVIEW Figure 19 shows a magnified view of the waveforms during13 of arc 15 clearly confirmed. The deviation time could be obtained from Figure 20, which is a magnified view (surrounded by a green ellipsoid in Figure 18), in which a distortion of the voltage waveform is of the waveforms surrounded by a yellow ellipsoid in Figure 19. The experimental voltage is drawn theclearly onset confirmed. of voltage The distortion. deviation Figure time could 19 show be obtaineds a magnified from Figure view of20, the which waveforms is a magnified during view arc in red, while the checking voltage and the range of checking voltage ±6 V are in green and light (surroundedof the waveforms by a surroundedgreen ellipsoid by ain yellow Figure ellipsoid 18), in inwhich Figure a distortion19. The experimental of the voltage voltage waveform is drawn is green,Energies 2020respectively., 13, 681 The arrows in Figure 20 indicate the durations when the experimental voltage13 of 15 clearlyin red, confirmed.while the checkingThe deviation voltage time and could the berange obtained of checking from Figure voltage 20, ±6 which V are is in a magnifiedgreen and viewlight exceeded the checking voltage ±6 V. The deviation time is 276 μs, which is longer than 70 μs. This ofgreen, the waveforms respectively. surrounded The arrows by in a yellowFigure ellipsoid20 indica inte Figurethe durations 19. The whenexperimental the experimental voltage is voltagedrawn case satisfies both of the criteria above, which implies that detecting the arc is possible. incheckingexceeded red, while voltagethe thechecking andchecking the voltage range voltage ±6 of V. checkingand The the deviation range voltage of time checking6 V is are 276 in voltageμ greens, which and±6 isV light longerare green,in thangreen respectively. 70 and μs. lightThis ± Thegreen,case arrows satisfies respectively. in both Figure of The the20 indicate arrowscriteria in theabove, Figure durations which 20 indica whenimplieste the thethat experimental durations detecting whenthe voltage arc the is exceeded possible.experimental the checking voltage voltageexceeded 6the V. Thechecking deviation voltage time ±6 is V. 276 Theµs, deviation which is time longer is than276 μ 70s, µwhichs. This is caselonger satisﬁes than 70 both μs. of This the ± casecriteria satisfies above, both which of the implies criteria that above, detecting which the implies arc is possible. that detecting the arc is possible.

Figure 18. Sample voltage and current waveforms at disconnection of the intact element conductor.

Figure 18. SampleSample voltage voltage and and current current waveforms waveforms at disc disconnectiononnectionArc extinction of the intact element conductor.

Arc extinction Arc ignition Arc extinction Arc ignition

Arc ignition

Figure 19. Magnified view of the part surrounded by a green ellipsoid in Figure 18.

Figure 19.19. MagniﬁedMagnified view of the part surrounded byby aa greengreen ellipsoidellipsoid inin FigureFigure 1818..

Figure 19. Magnified view of the part surrounded by a green ellipsoid in Figure 18.

Deviation Time

Deviation Time Figure 20. Magniﬁed view of theDeviation part surrounded Time by a yellow ellipsoid in Figure 19.

In the case of nonlinear loads, it is considered that the checking voltage is obtained by the same procedures as the case of a resistive load shown in Figure 14. It was conﬁrmed that the checking voltage was obtained for a vacuum cleaner. Further examination will be necessary using other nonlinear loads. Energies 2020, 13, 681 14 of 15

Acquisition of the checking voltage when an arc generates within a half cycle after the commencement of current supply is also a subject to be investigated.

5. Conclusions Ignition of combustibles related to an arc was studied, which were placed on two kinds of damaged power supply cords. The ﬁrst is caused by an arc at the time of a short circuit, which occurs when broken element conductors in one cable assume contact with conductors in the other cable at the damaged part of a power supply cord. Here, shrinkage stress acting on PVC during the cooling process may be a possible cause. The second is caused by an arc observed at the time of breakage of a remaining intact element conductor by Joule heating. In both cases, combustibles probabilistically ignite when the temperature of a minute portion of them may exceeds the ﬂash point by contact with arc and/or a hot piece of melted element conductor. A possible feature, deviation time, which is obtained by comparing distorted voltage waveform at arc with the checking voltage waveform, can potentially be used to detect arc in the two scenarios. On this note, further investigations are required to verify the proposed ignition mechanisms of combustibles and to evaluate validity of the feature in other situations.

Author Contributions: Conceptualization, K.T., Y.I., Y.M., and W.L.; methodology, K.T., Y.I., and W.L.; validation, K.T.; formal analysis, K.T.; investigation, K.T. and Y.I; resources, K.T. and Y.I.; data curation, K.T. and Y.I.; writing—original draft preparation, K.T.; writing—review and editing, Y.I., Y.M., and W.L.; visualization, K.T.; supervision, Y.M.; project administration, Y.M.; funding acquisition, Y.M. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by JSPS KAKENHI Grant Number JP17K06299. Acknowledgments: The authors would like to thank Japan Electric Cable Technology Center for the preparation of the samples by performing the bending test of the power supply cords. Conﬂicts of Interest: The authors declare no conﬂict of interest.

References

1. Tokyo Fire Department. Survey of Fires Accidents in 2019. Available online: http://www.tfd.metro.tokyo.jp/ hp-cyousaka/kasaijittai/h31/ (accessed on 18 December 2019). 2. Committee for Investigating Electric Fire Prevention Methods at a Big Earthquake. Report of Investigation and Promotion of Electric Fire Prevention Methods at a Big Earthquake. 2015, pp. 11–13. Available online: http://www.bousai.go.jp/jishin/syuto/denkikasaitaisaku/pdf/guideline_houkoku.pdf (accessed on 18 December 2019). 3. Gregory, G.D.; Wong, K.; Dvorak, R.F. More About Arc-Fault Circuit Interrupters. Ieee Trans. Ind. Appl. 2004, 40, 1006–1011. [CrossRef] 4. Abe, T.; Fukagawa, K.; Mizuno, Y.; Yoshida, A. Arc Discharge Detection Caused by Short-Circuit in Ac Power Supply Cord. In Proceedings of the IEEE Conference on Electrical Insulation and Dielectric Phenomena, Toronto, Canada, 16–19 October 2016. Paper No. 4B-21. 5. UL 1699. Standard for Safety, Arc-Fault Circuit-Interrupters, 3rd ed.; UL: Bensenville, IL, USA, 2017. 6. IEC 62606. General Requirements for Arc Fault Detection Devices, 1st ed.; International Electrotechnical Commission: Geneva, Switzerland, 2017. 7. Shea, J.J. Comparing 240 Vrms to 120 Vrms Series Arcing Faults in Residential Wire. In Proceedings of the 54th IEEE Holm Conference on Electrical Contacts, Orlando, FL, USA, 27–29 October 2008; pp. 218–224. 8. Shea, J.J. Identifying Causes for Certain Types of Electrically Initiated Fires in Residential Circuit. Fire Mater. 2011, 35, 19–42. [CrossRef] 9. Kato, K.; Kataoka, M. Study of the Electric Fire Mechanism Caused by the Fault of Electric Cords for Domestic Use. Rep. Fire Technol. Saf. Lab. 2000, 37, 144–150. 10. Awart, J.; Loud, J.; Slee, D. Arcing and Fire–Case Studies. In Proceedings of the IEEE Symposium on Product Compliance Engineering, Toronto, ON, Canada, 26–28 October 2009. 11. Shea, J.J. Conditions for Series Arcing Phenomena in PVC Wiring. IEEE Trans. Compon. Packag. Technol. 2007, 30, 532–539. [CrossRef] Energies 2020, 13, 681 15 of 15

12. Ishikawa, Y.; Takenaka, K.; Mizuno, Y.; Yoshida, A. Discharge-Induced Ignition of Combustibles on Ac Power-Supply Cords. In Proceedings of the IEEE Conference on Electrical Insulation and Dielectric Phenomena, Cancun, Mexico, 21–24 October 2018. Paper No. 6A-13. 13. Restrepo, C.; Staley, P.S. Systems, Devices, and Methods for Arc Fault Detection. US Patent 0252603 A1, 1 November 2007. 14. Ji, H.-K.; Wang, G.; Kim, W.-H.; Kil, G.-S. Optimal Design of a Band Pass Filter and an Algorithm for Series Arc Detection. Energies 2018, 11, 992. [CrossRef] 15. Wu, C.J.; Liu, Y.W. Smart detection technology of serial arc fault on low-voltage indoor power lines. Int. J. Electr. Power Energy Syst. 2015, 69, 391–398. [CrossRef] 16. Yang, K.; Zhang, R.; Yang, J.; Liu, C.; Chen, S.; Zhang, F. A Novel Arc Fault Detector for Early Detection of Electrical Fires. Sensors 2016, 16, 500. [CrossRef][PubMed] 17. Parise, G.; Parise, L. Unprotected Faults of Electrical and Extension Cords in AC and DC Systems. IEEE Trans. Ind. Appl. 2014, 50, 4–9. [CrossRef] 18. Moon, W.-S.; Kim, J.-C.; Jo, A.; Bang, S.-B. Ignition Characteristics of Residential Series Arc Faults in 220-V HIV Wires. IEEE Trans. Ind. Appl. 2015, 51, 2054–2059. [CrossRef] 19. JIS C 3005. Test Methods for Rubber or Plastic Insulated Wires and Cables; Japanese Standards Association: Tokyo, Japan, 2014; Section 4.27. 20. Polymer Properties Database. Available online: http://polymerdatabase.com/polymers/polyvinylchloride. html (accessed on 18 December 2019). 21. Takenaka, K.; Mizuno, Y.; Yoshida, A. Condition Monitoring of Damaged AC Power Supply Cord Using Voltage Waveform. In Proceedings of the International Conference on Condition Monitoring and Diagnosis (CMD2018), Perth, Australia, 23–26 September 2018. Paper No. P.47.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).