sensors Article

Use of Distributed Temperature Sensing Technology to Characterize Fire Behavior Douglas Cram 1, *, Christine E. Hatch 2 , Scott Tyler 3 and Carlos Ochoa 4 1 2 3 4

*

Extension Animal Sciences and Natural Resources, New Mexico State University, Las Cruces, NM 88003, USA Department of Geosciences, University of Massachusetts, Amherst, MA 01003, USA; [email protected] Geological Sciences and Engineering, University of Nevada, Reno, NV 89557, USA; [email protected] Department of Animal and Rangeland Sciences, Oregon State University, Corvallis, OR 97331, USA; [email protected] Correspondence: [email protected]; Tel.: +1-575-646-8130

Academic Editor: Vittorio M. N. Passaro Received: 31 August 2016; Accepted: 9 October 2016; Published: 17 October 2016

Abstract: We evaluated the potential of a fiber optic cable connected to distributed temperature sensing (DTS) technology to withstand wildland fire conditions and quantify fire behavior parameters. We used a custom-made ‘fire cable’ consisting of three optical fibers coated with three different materials—acrylate, copper and polyimide. The 150-m cable was deployed in grasslands and burned in three prescribed fires. The DTS system recorded fire cable output every three seconds and integrated temperatures every 50.6 cm. Results indicated the fire cable was physically capable of withstanding repeated rugged use. Fiber coating materials withstood temperatures up to 422 ◦ C. Changes in fiber attenuation following fire were near zero (−0.81 to 0.12 dB/km) indicating essentially no change in light gain or loss as a function of distance or fire intensity over the length of the fire cable. Results indicated fire cable and DTS technology have potential to quantify fire environment parameters such as heat duration and rate of spread but additional experimentation and analysis are required to determine efficacy and response times. This study adds understanding of DTS and fire cable technology as a potential new method for characterizing fire behavior parameters at greater temporal and spatial scales. Keywords: fiber-optic; FO-DTS; prescribed fire; rangeland; spatial mapping

1. Introduction Fire science, effects, and modeling research require measures of fire behavior parameters such as fireline intensity, heat duration, and rates of spread [1,2]. Quantifying these parameters is challenging in the laboratory and notoriously difficult in the field where instrument limitations, variations in fuel characteristics, and the turbulent nature of fluid dynamics and combustion lead to drastic variations over time and space. Options for quantifying these parameters include remotely sensed data from instruments mounted on aircraft or satellites with large pixel footprints or ground-based sensors within or above the flames [2]. Spatial limitations are a significant shortcoming of current ground-based instruments when used to quantify fire behavior parameters because they operate at an individual point scale (e.g., thermocouples) [3,4]. Filling the gap in the spatial domain as well as the temporal scale can be critical to improving our understanding of fire behavior in wildland fuels. We tested a custom built fiber-optic ‘fire cable’ using distributed temperature sensing (DTS) technology under wildland fire conditions. Distributed temperature sensing technology has become a widely-used environmental monitoring tool in recent years ranging from applications in hydrology [5], atmospheric science [6] and oceanography [7]. Advantages of DTS technology include the opportunity

Sensors 2016, 16, 1712; doi:10.3390/s16101712

www.mdpi.com/journal/sensors

Sensors 2016, 16, 1712

2 of 12

to record data at a high level of precision and at high spatial and temporal frequencies in air, water, soil or other media. Distributed temperature sensing technology is also widely used for industrial [8] and transportation (e.g., tunnel) fire detection [9]. In fire detection applications, DTS is used primarily to indicate presence or absence of fire rather than continuously monitoring fire behavior parameters. Distributed temperature sensing technology has not been used in natural settings for fire characterization nor has it been used for prescribed fire applications. Distributed temperature sensing technology is based on sending (launching) a light pulse down an optical fiber and analyzing the return signal of backscattered Raman-spectra light at specific frequencies that are either sensitive (anti-Stokes) or not sensitive (Stokes) to changes in temperature. Photons returning within specific frequency bands are recorded in timed ‘bins’ which, based on the speed of light, represent a discrete length of cable a known distance from the detector. After calibration to locations of known temperature along the cable, temperatures can be resolved for the entire length of the cable at the desired time interval. Typical commercial DTS systems provide an integrated temperature every 1–2 m of fiber at frequencies from 0.01 to 1 Hz. We conservatively estimated and reported temperature resolutions to ~0.1 ◦ C as achieved in the field. Distributed temperature sensing systems are capable of measuring temperatures as high as 600 ◦ C and since the optical fiber is glass, high temperatures can be tolerated. However, repeated and prolonged exposure to high temperatures (i.e., >400 ◦ C) promotes or accelerates ‘fiber darkening’ a product of hydrogen contamination, especially in hydrogen-saturated environments such as petrochemical applications [10,11]. The presence of free hydrogen and other weathering products results in greater adsorption of incident light over time, resulting in less light transmitted, and less light returned to detectors. Typical optical fiber used for telecommunication is designed to have insignificant darkening at temperatures between 60 ◦ C and 85 ◦ C over the life of operation (e.g., tens of years). At temperatures above this design threshold, darkening may occur more rapidly, resulting in time-dependent light transmission/scattering behavior. To overcome this issue, fibers can be coated with a variety of materials including gold [12] and then surrounded by a hydrogen scavenging gel. In this work, we tested several different fiber coatings to determine how they performed following multiple prescribed fires. Additional fiber challenges related to wildland conditions are as follows: the optical fiber is fragile (e.g., typical diameter is 125 microns) and must be protected from strain and abrasion. Typically, strength elements are designed to mitigate longitudinal strain, stainless steel capillary tubes protect the fiber from radial strain, and plastic jacketing is used to protect the fiber optic cable from ultraviolet radiation. In this work, we consider the fiber-optic fire cable to include optical fibers (i.e., three fibers), coating material on the optical fibers (three different materials were used), and strength elements (the cable used in this study did not have a protective jacket). The objective of this study was to evaluate the potential of a fiber optic cable connected to DTS technology to withstand the fire environment (i.e., exposure to repeated combustion temperatures and deployment abuse) and quantify fire behavior parameters such as temperature, heat duration and rate of spread. 2. Experimental Section 2.1. Study Site Our experiment site was located on the Corona Range and Livestock Research Center, Torrance County, NM, USA (UTM 13 S 463655 E 3793075 N) at 1870 m above sea level. The ranch is operated by New Mexico State University for range and animal science research and teaching purposes. The burn site vegetation and fuel was characterized as short-grass blue grama (Bouteloua gracilis) grassland with little to no shrubs outside of an occasional cholla cactus (Cylindropuntia imbricata). Average annual precipitation between 1990 and 2007 was 330 mm, with most of the rain falling between July and September. Average daily maximum air temperature is 10 ◦ C during winter and 29 ◦ C in summer. Average daily minimum air temperature is −3.9 ◦ C in winter and 13.9 ◦ C in summer [13].

Sensors 2016, 16, 1712 Sensors 2016, 16, 1712

3 of 12 3 of 13

2.2. Study Design Three prescribed fires were implemented to test the applicability of the DTS technology to measuring fire fireparameters parameters such as duration (i.e., residence heat rate and of spread, and such as duration (i.e., residence heat time), ratetime), of spread, temperature. temperature. The prescribed burns were conducted on three different 20 by 27 m plots on 15 March The prescribed burns were conducted on three different 20 by 27 m plots on 15 March 2012. Vegetation −1 one 2012.senescent Vegetation was senescent the time offuel burn. Herbaceous burn plot was at the time of burn.atHerbaceous loading for burnfuel plotloading one wasfor 733 kg·ha (136 was SE), − 1 −1 −1 733 andfor 706burn kg ha SE)No forfuel burn plot two. fuel loading wasplot recorded and kg·ha 706 kg·(136 ha SE), (66 SE) plot(66 two. loading wasNo recorded for burn three. for Theburn size plot The size of the rectangular burn plots by astudy priorat prescribed fire study at of thethree. rectangular burn plots was established by a was priorestablished prescribed fire the same location [14]. the6.7-m samefirebreak locationwas [14].installed A 6.7-monfirebreak was of installed onusing all four sides of the plots using a were road A all four sides the plots a road grader. The three burns grader. three12:56, burnsand were ignited at 11:20, 12:56, and 13:52 h, respectively, each burn lasting ignited The at 11:20, 13:52 h, respectively, with each burn lasting aboutwith 8 min. Following the about 8 min. the burn, fire cable wasignition movedwas to astarted new plot. ignition was started burn, the fire Following cable was moved to athe new plot. Each withEach a backfire, followed by a with backfire, followed bya ahead flank and 1). finished with fireas(Figure 1). A head firethe is flank afire, and finished with firefire, (Figure A head fire ais head defined a fire spreading with defined as a fire spreading with the wind. A back fire is defined as a fire spreading against the wind. wind. A back fire is defined as a fire spreading against the wind. A flank fire is defined fire spreading A flank fire is defined fire spreading perpendicular to the wind. perpendicular to the wind.

Figure 1. Photograph Photograph of of prescribed prescribed fire fire (plot (plot 1) 1) at at Corona Corona Range Range and and Livestock Livestock Research Research Center, Center, Torrance County, County, New New Mexico Mexico showing showing backfire backfire (visible (visible on on the left), flank fire (foreground) and head Torrance fire (moving from right to left). The ‘fire cable’ is visible visible on on the the ground. ground.

2.3. 2.3. DTS DTS Technology Technology Standard telecommunication telecommunication fiber-optic not designed designed to to withstand withstand long long periods periods at at Standard fiber-optic cables cables are are not combustion temperaturesgenerated generated a result of burning wildland fuels. test the combustion temperatures as aasresult of burning wildland fuels. In orderInto order test thetofeasibility feasibility of using DTS technology for wildland fire applications, a 150-m specialized fiber-optic fire of using DTS technology for wildland fire applications, a 150-m specialized fiber-optic fire cable was cable was custom made The fire cable consisted of three optical fibers with three different custom made (Figure 2).(Figure The fire2).cable consisted of three optical fibers with three different coatings coatings ranging from acoating low-cost coating typical of telecommunication to specialty ranging from a low-cost typical of telecommunication applicationsapplications to specialty materials to materials to protect the optical fibers. Each fiber was armored in a stainless steel capillary tube and protect the optical fibers. Each fiber was armored in a stainless steel capillary tube and backfilled backfilled inert gas. Optical glass is sufficiently heat so resistant, so multi-mode standard multi-mode 50-μm with inert with gas. Optical glass is sufficiently heat resistant, standard 50-µm core and core and 125-μm cladding fiber was used. Typical acrylate buffer coating used on most optical fibers 125-µm cladding fiber was used. Typical acrylate buffer coating used on most optical fibers is rated ◦ C.120 is to withstand only We included fiber coated acrylate to test its performance in to rated withstand only 120 We°C. included a fibera coated with with acrylate to test its performance in an an extreme environment. addition,we weincluded includedtwo twomore morespecialized specializedfibers, fibers, each each coated coated with with aa extreme environment. InIn addition, different heat-resistant material: copper and polyimide. Each fiber was encased in an individual,

Sensors 2016, 16, 1712

4 of 12

Sensors 2016, 16, 1712

4 of 13

different heat-resistant material: copper and polyimide. Each fiber was encased in an individual, rigid rigid stainless steeland tube and overstuffed fiberislength ~0.1%than longer than tube the steel tube stainless steel tube overstuffed (i.e., the(i.e., fiberthe length ~0.1% is longer the steel length) to length) to accommodate bending. Typically these tubes would be backfilled with a hydrophobic gel accommodate bending. Typically these tubes would be backfilled with a hydrophobic gel to eliminate to eliminate water, but since the gels havetobeen known to expand flow when out ofheated, the tubes water, but since the gels have been known expand and flow out of and the tubes the when tubes heated, tubes were filled during with gel. Instead, during prescribed fire amount operations a small amount were notthe filled with gel.not Instead, prescribed fire operations a small of dry nitrogen gas of dry nitrogen gas was continuously pumped through all three stainless-steel tubes to protect the was continuously pumped through all three stainless-steel tubes to protect the fibers, prevent air or fibers, prevent air or water from entering, and reduce condensation occurring inside the tubes. water from entering, and reduce condensation from occurring insidefrom the tubes. Three stainless-steel Three stainless-steel tubes were braided together around a single steel strength member to the tubes were braided together around a single steel strength member to form the finished form armored finished armored cable. The resulting braided bundle had a finished diameter of 9 mm. Typically, cable. The resulting braided bundle had a finished diameter of 9 mm. Typically, cables are encased cables are encased in UV-resistant other protective plastic,the butthermal to maximize the thermal in UV-resistant or other protectiveor plastic, but to maximize response, we left response, the metal we left the metal bundle without additional coating. Since the predominant mode of heat transport bundle without additional coating. Since the predominant mode of heat transport for sensing will be for sensing will be conductive, we expect the dominantly metal cable bundle to respond rapidly conductive, we expect the dominantly metal cable bundle to respond rapidly (seconds) to changes (seconds) totemperature. changes in ambient temperature. It should also be noted that are likely contact in ambient It should also be noted that fibers are likely in fibers contact with theinstainless with the stainless steel walls of the capillary tube due to their helical arrangement inside the steel walls of the capillary tube due to their helical arrangement inside the tube (generated by tube ~1% (generated by ~1% over-stuffing) and of due to the absence over-stuffing) and due to the absence hydrophobic gel.of hydrophobic gel.

Figure Figure 2. 2. Cross-section Cross-section schematic schematic of of custom-made custom-made fiber fiber optic optic ‘fire ‘fire cable’. Fire Fire cable cable consisted consisted of three hollow hollow stainless stainless steel steel capillary capillary tubes tubes twisted twisted together together around around a central strength element. Each Each tube contained contained one one 50/125-μm 50/125-µm multi-mode multi-mode optical optical fiber. fiber. Each Each fiber fiber had had aa different different coating: coating: copper, copper, polyimide hydrogen scavenging gel gel waswas usedused for this As such, to the polyimide and andacrylate. acrylate.No No hydrogen scavenging for cable. this cable. As contrary such, contrary figure illustration, it is likely the optical fibers were in contact with the to the figure illustration, it is likely the optical fibers were in contact with thestainless stainlesssteel steelwall wallof of the the capillary tubes tubes for for much much of of the the length length of the cable. Nitrogen Nitrogen gas gas was pumped through the stainless capillary steel tubes tubes during during prescribed prescribed fires. Fire cable was constructed by AFL Telecommunications. steel Telecommunications.

In order totoevaluate evaluate physical performance of fiber each coating fiber coating in cable the fire cable we In order thethe physical performance of each in the fire we compared compared signaland strength and light propagation of one Raman frequency (i.e.,before Stokes) before the signal the strength light propagation of one Raman frequency (i.e., Stokes) and afterand the after the prescribed burns. The incident laser light used by our DTS was 1040 nm, which is Raman prescribed burns. The incident laser light used by our DTS was 1040 nm, which is Raman backscattered backscattered into Stokes and anti-Stokesofwavelengths of1000 1080nm, nm respectively. and 1000 nm,We respectively. We into Stokes and anti-Stokes wavelengths 1080 nm and used the Stokes used the Stokes backscattered light for1080 thisnm). analysis (i.e., 1080 nm). Because therelight is variability backscattered light for this analysis (i.e., Because there is variability between pulses we between light pulses we averaged 50 traces before the cable was subjected to the burn and 50 traces averaged 50 traces before the cable was subjected to the burn and 50 traces after the burn. The cable after the burn. The cable was not disconnected or impinged between trace sampling. As a result, we was not disconnected or impinged between trace sampling. As a result, we assumed the only change assumed the only between two was averages offire light intensity was dueFor to fire the between these twochange averages of lightthese intensity due to effects on the cable. eacheffects meteron along cable. For each meter along the cable, we calculated the amplitude of light intensity lost (in the cable, we calculated the amplitude of light intensity lost (in decibels/km) since its ‘launch’ into the decibels/km) since its ‘launch’ into the fiber from the DTS unit for Stokes frequency. We then fiber from the DTS unit for Stokes frequency. We then subtracted the averages (before minus after) to subtracted the due averages (before after) assess change to heating onminus the cable fortoallassess three change fibers. due to heating on the cable for all three fibers.

Sensors 2016, 16, 1712 Sensors 2016, 16, 1712

5 of 13 5 of 12

2.4. Field Data Collection 2.4. Field Data Collection We deployed the fire cable within each burn unit and arranged it to maximize spatial coverage of theWe experimental Eachwithin opticaleach fiberburn wasunit connected to oneitchannel of anspatial Ultimacoverage SR (Silixa deployed theplots. fire cable and arranged to maximize of Ltd., Hertfordshire, UK)Each DTS optical unit housed inside a closed to canopy (2.1 by 4.2 m)Ultima trailer SR for (Silixa protection. the experimental plots. fiber was connected one channel of an Ltd., We used a portable generator to provide forcanopy the electronic equipment the DTS Hertfordshire, UK) DTS unit housed insidepower a closed (2.1 by 4.2 m) trailerincluding for protection. We unit. used Within thegenerator trailer, two 75 L plastic coolers were filled with 37 L ofincluding warm water and cold a portable to provide power for the electronic equipment the DTS unit. water Withinplus the ice, respectively to create a ‘warm’ or ambient andwater a ‘cold’ used plus for DTS calibration trailer, two 75 L plastic coolers were filled with 37 L bath of warm and bath cold water ice, respectively purposes. While the DTS unit weand used had abath default we opted to preform to create a ‘warm’ or ambient bath a ‘cold’ used calibration, for DTS calibration purposes. Whileour the own DTS calibration test to aensure collectedwe accurate data that related actual temperatures. Four unit we used had defaultwe calibration, opted to preform our owntocalibration test to ensure we eight-meter coils of thethat fire related cable were placed in the warmFour and eight-meter cold baths (i.e., coils eachwere end collected accurate data to actual temperatures. coilstwo of the fireatcable of the fire cable). Theand first andbaths last meter of calibration coil data were discarded, leaving six last meters of placed in the warm cold (i.e., two coils at each end of the fire cable). The first and meter cable for calibration at each of the cable in each bath. of The Ultima DTS has a minimum of calibration coil data were end discarded, leaving six meters cable for calibration at each endsample of the resolution of bath. 12.5 The cm. Ultima For thisDTS installation, the sampling resolutionofwas to installation, 25.3 cm to cable in each has a minimum sample resolution 12.5increased cm. For this improve temperature resolution. Typically, thecmability to resolve differences in temperature along the sampling resolution was increased to 25.3 to improve temperature resolution. Typically, the the fiber (spatialdifferences resolution)inistemperature approximately the (spatial sampling resolution. However, fortwice this ability to resolve alongtwice the fiber resolution) is approximately experiment, calculated spatialforresolution closer we to calculated 1 meter spatial [15]. Platinum-resistance the sampling we resolution. However, this experiment, resolution closer to ◦ thermometers (PT100) rated tothermometers 0.1 °C accuracy were rated placed bath andwere temperatures 1 m [15]. Platinum-resistance (PT100) to in 0.1each C accuracy placed in were each recorded the DTS unit at the same three-second intervals DTS data from the fibers. as In DTS addition, bath and by temperatures were recorded by the DTS unit at the as same three-second intervals data independent of fire temperature using three, type-K from the fibers.measurements In addition, independent measurementswere of fire obtained temperature were obtained using (Chromel-Alumel), thermocouplesthermocouples that were deployed during each of the three three, type-K (Chromel-Alumel), that were deployed during eachburn of thetrials. threeThree burn thermocouples were positioned within 0.2 cm of the fire cable at three known locations (i.e., 55, 100, trials. Three thermocouples were positioned within 0.2 cm of the fire cable at three known locations and along (Figure 3). These locations corresponded with head, back, and flankback, fire (i.e., 123 55, m 100, and fire 123 cable) m along fire cable) (Figure 3). These locations corresponded with head, conditions during burn. Head, back and flank were follows along cable:along 30–73the m, 85– and flank fire conditions during burn. Head,fires back andasflank fires werethe asfire follows fire 112 m, and 118–130 m, respectively. Thermocouples were connected to a CR10X datalogger cable: 30–73 m, 85–112 m, and 118–130 m, respectively. Thermocouples were connected to a CR10X (Campbell Inc., Logan, Inc., UT, USA) that was programmed to record maximum temperature, dataloggerScientific (Campbell Scientific Logan, UT, USA) that was programmed to record maximum time of occurrence, and average temperature at five-second intervals. Figure 3 shows the temperature, time of occurrence, and average temperature at five-second intervals. Figure 3placement shows the of the fire cable To quantifyTo the total amount of amount thermal of energy (TTE) measured placement of theand firethermocouples. cable and thermocouples. quantify the total thermal energy (TTE) at a knownatlocation onlocation each instrument we calculated the sum ofthe measured energy over time over (i.e., measured a known on each instrument we calculated sum of measured energy the duration of the prescribed fire). time (i.e., the duration of the prescribed fire).

Figure 3. Schematic Schematic layout layout of 150-m fiber-optic ‘fire cable’ in experimental burn plots. This layout was repeated in plots 1, 2, and 3. Each grassland grassland plot plot was was 20 by 27 m. The The fire fire cable cable runs from the distributed distributed temperature temperature sensing sensing (DTS) (DTS) unit unit into into aa cold cold calibration calibration bath, bath, into into aa hot hot calibration calibration bath, bath, out into the plot and back through both baths. The The DTS DTS unit unit and and baths baths were housed housed in a protective linear transects approximately 8 m8apart. Ignition order was trailer. The The fire firecable cableruns runsinintwo tworoughly roughly linear transects approximately m apart. Ignition order a follows: backback fire, fire, flankflank fire, head fire. fire. was a follows: fire, head

Sensors 2016, 16, 1712

6 of 12

3. Results and Discussion Fire cable and DTS technology detected short-grass wildland fire combustion without catastrophic failure during three prescribed fires. This was significant, because to our knowledge, this specific application had not before been tested in wildland fire conditions. Results indicated the fire cable was physically capable of withstanding repeated rugged use. Also of interest was how the three fiber coatings affected attenuation. Changes in optical fiber attenuation following fire were near zero (−0.81 to 0.12 dB/km) for the polyimide and acrylate coated fibers indicating essentially no change in light gain or loss as a function of distance, fire intensity, or maximum temperature reached following three prescribed burns (Table 1). The polyimide and acrylate coated fibers showed a relatively consistent value for attenuation along the entire length of the fire cable (Figure 4). As for the copper coated fiber, while the attenuation signature was logarithmic in shape—which is unusual for optical fibers—the change in attenuation was minor (Figure 4, Table 1). We believe the logarithmic shape of the attenuation signal was an artifact of the manufacturing process (e.g., bend memory) as this condition existed before prescribed burns were conducted. In order to characterize changes in attenuation on this fiber, three sections were selected with different representative slopes for attenuation calculations. Results showed changes from −13.35 to 9.53 dB/km after the prescribed burns (Table 1). We would expect an increase in attenuation if the burn caused any damage to the fiber, yet in two sections of the fiber we observe the opposite. This decreased attenuation may be the result of relaxation of manufacturing effects such as bend memory, but is not likely to have been induced by the burn specifically. Hypothetically, the changing slope along the entire length of the copper fiber could result in increased noise or less accurate temperature measurements further down the cable but this was not supported by the data. The anomalous attenuation curve observed in this particular copper-coated fiber may not be representative of all copper-coated fibers. A microscope examination of the glass fibers following decommission would reveal more about the nature and extent of any possible fiber or coating damage. Table 1. Losses in Raman backscattered amplitude (Stokes frequency) following three prescribed burns. Before Burn

After Burn

Change b,c

DTS Fiber Coating

Loss (dB)

Length (m) a

Attenuation (dB/km)

Loss (dB)

Attenuation (dB/km)

Attenuation (dB/km)

Polyimide Copper

0.22 1.49 2.63 2.96 0.26

100.7 11.8 33.0 95.8 121.2

2.1 126.8 79.7 30.9 2.2

0.13 1.60 2.19 2.11 0.28

1.9 135.8 69.3 24.5 2.3

−0.81 9.53 −13.35 −8.78 0.12

Acrylate a

Attenuation was calculated before and after burns over a length of cable with representative slope. For the polyamide-coated fiber, loss was calculated from 20.6 to 121.2 m. For the copper-coated fiber where attenuation was changing, three representative sections were chosen from 3.2 to 15.0 m, 18.0 to 51.1 m, and 56.0 to 151.7 m. Acrylate losses were calculated from 20.5 to 141.7 m; b Calculated as average attenuation (dB/km) of the last 50 traces after burning plot 3 minus average attenuation of the first 50 traces before burn in plot 1; positive values indicate an increase in attenuation; negative values indicate a decrease in attenuation; c Fibers with consistent attenuation (i.e. relatively constant slope of Stokes intensity vs. distance) have near-zero change in attenuation (i.e., polyimide, acrylate). Large decreases in attenuation are most likely a relaxation of fabrication-induced attenuation, rather than any effect of fiber burning, which would be expected to increase attenuation if damage to fibers occurred.

Differences in performance given three different fiber coatings are illustrated in Figure 5, which shows the maximum recorded temperature during the burn vs. distance along the cable for all three burn plots.

Sensors 2016, 16, 1712 Sensors 2016, 16, 1712 Sensors 2016, 16, 1712

7 of 13 7 of 12 7 of 13

Figure 4. 4.Attenuation Attenuation (dB/km) between raw DTS DTS data data (backscattered (backscatteredRaman RamanStokes Stokesfrequency frequency data (backscattered Raman Stokes Figure Attenuation(dB/km) (dB/km) between between raw raw DTS amplitude) before (blue) and after (red) each burn in plot 1, 2, and 3 for (a) Polyimide coated fiber; amplitude) before (blue) 1, 2, 2, and and33for for(a) (a)Polyimide Polyimidecoated coated fiber; (b) amplitude) before (blue)and andafter after(red) (red)each each burn burn in plot 1, fiber; (b) (b)Copper Copper coated fiber; and Acrylate coated fiber.The The first traces theStokes Stokes signals coated fiber; and (c)(c) Acrylate coated fiber. first 50 traces ofofof the signals areare Copper coated fiber; and (c) Acrylate coated fiber. The first 5050 traces the Stokes signals are averaged together, and the last 50 traces after each burn burn are areaveraged averagedtogether togetherand and plotted together. averaged together,and andthe thelast last50 50traces tracesafter after each plotted together. averaged together, are averaged together and plotted together. There is little to no change in attenuation after each burn. There is little to no change in attenuation after each burn. There is little to no change in attenuation after each burn.

(a) (a)

(b)

(b)

(c) (c)

Figure 5. Maximum fire cable temperatures as recorded by DTS technology by distance (m) for polyimide (thicker, orange), copper (dashed purple), and acrylate (green) optical fiber coatings fromfor Figure 5. Maximum Figure 5. Maximum fire fire cable cable temperatures temperatures as as recorded recorded by by DTS DTS technology technology by by distance distance (m) (m) for (a) Plot 1 (Max = 421.9 °C on copper); (b) Plot 2 (Max = 223.2 °C on copper); and (c) Plotfrom 3 polyimide (thicker, orange), copper (dashed purple), and acrylate (green) optical fiber coatings polyimide (thicker, orange), copper (dashed purple), and acrylate (green) optical fiber coatings (Max = 395.2 °C=on421.9 copper). Calibration bath were as follows: ice baths at 7–15and m and 146– (a) Plot (Max °C on (b) locations PlotPlot 2 (Max = = 223.2 3 ◦ C copper); from (a) 1Plot 1 (Max = 421.9 on copper); (b) 2 (Max 223.2°C◦ Cononcopper); copper); and (c) (c) Plot Plot 3 154=m, and °C ambient temperature baths at 16–24 m and 137–145 m. Head, back and flank fires were as (Max 395.2 on copper). Calibration bath locations were as follows: ice baths at 7–15 m and 146– ◦ (Max = 395.2 C on copper). Calibration bath locations were as follows: ice baths at 7–15 m and follows: 30–73 m, 85–112 m, andbaths 118–130 m, respectively. Fire cable that was outside the fire linewere was as 154 m, and ambient temperature at 16–24 m and 137–145 m. Head, back andand flank fires 146–154 m, and ambient temperature baths at 16–24 m and 137–145 m. Head, back flank fires were as follows: 0–29 m, 74–84 m, 113–117, and 131–154 m. follows: 30–73 m, 85–112 m, and 118–130 m, respectively. Fire Fire cable that that was was outside the fire as follows: 30–73 m, 85–112 m, and 118–130 m, respectively. cable outside theline fire was line as follows: 0–29 m, 74–84 m, 113–117, and 131–154 m. was as follows: 0–29 m, 74–84 m, 113–117, and 131–154 m.

Thermocouple 366.7 248.1 344.0 DTS fiber coating polyimide 122.7 173.1 112.3 307.1 104 copper 166.1 202.3 98.5 421.9 105 acrylate 169.1 145.7 143.2 310.1 106 Plot 2 2016, 16, 1712 Sensors 8 of 12 Thermocouple 184.4 363.2 144.1 DTS fiber coating polyimide 91.7 97.6 100.1 178.1 118 As predicted, because of its relatively lower heat capacity, the copper coated fiber recorded greater copper 105.1 141.7 132.4 223.2 118 or near-equal maximum temperatures100.8 when compared and acrylate coated119 fibers acrylate 82.81 122.2to the polyimide178.3 at head, back and flank fire locations. The polyimide and acrylate coated fibers recorded similar Plot 3 Thermocoupleacross all380.7 239.6 324.0 and flank). Absolute peak temperatures temperatures fire conditions (i.e., head, back, DTS fiber coating recorded by DTS technology were 407, 422 and 310 ◦ C on polyimide, copper and acrylate coated fibers. polyimide 108.9 97.1 243.4 243.4 123 These peak temperatures were recorded locations in plot one. copper 147.9 105.4at back fire170.6 395.2Fire cable temperatures 124 derived from DTS technology were considerably lower acrylate 113.6 121.4 94.0 than thermocouple 254.2 temperatures at 125head,

back aand flank fire along locations (i.e., meter 100, maximum and 123) attemperature all three plots 6, Table 2). The Meter mark fiber-optic cable55, where was(Figure measured by DTS and heat c capacity of the firebcable may hinder its ability to accurately capture peakoftemperatures, especially Maximum temperature as measured along entire length fire cable by DTS; thermocouple; whenLocation subjected to the rapid changes associated fire conditions (Figure 6). of maximum temperature as measuredwith alonghead entire length of fire cable by DTS.

Figure Fire cable temperatures as recorded by DTS (dashed technology lines) and Figure 6.6.Fire cable temperatures (◦ C) as(C) recorded by DTS technology lines)(dashed and thermocouples thermocouples (solid lines) during three prescribed fire at distances of 55, 100, and 123 m along the (solid lines) during three prescribed fire at distances of 55, 100, and 123 m along the cable. Data from cable. Data from(i.e., the plots three 1,burns (i.e., 2, and from 3) were recorded fibers with the three burns 2, and 3) plots were 1, recorded optical fibersfrom with optical polyimide (a,d,g), polyimide (a,d,g), copper (b,e,h), and acrylate (c,f,i) coatings. Head (55 m), back (100 m) and flank copper (b,e,h), and acrylate (c,f,i) coatings. Head (55 m), back (100 m) and flank (123 m) fires were (123 fires were usedm) to burn plots.used to burn plots. Table 2. Maximum temperature as measured by thermocouple and DTS fire cable during three prescribed burns. Location and Sensor

T (◦ C) at 55 m a

T (◦ C) at 100 m

T (◦ C) at 123 m

366.7

248.1

344.0

122.7 166.1 169.1

173.1 202.3 145.7

112.3 98.5 143.2

184.4

363.2

144.1

91.7 105.1 82.81

97.6 141.7 100.8

100.1 132.4 122.2

380.7

239.6

324.0

108.9 147.9 113.6

97.1 105.4 121.4

243.4 170.6 94.0

Absolute Maximum T (◦ C) b

Location (m) c

307.1 421.9 310.1

104 105 106

178.1 223.2 178.3

118 118 119

243.4 395.2 254.2

123 124 125

Plot 1 Thermocouple DTS fiber coating polyimide copper acrylate Plot 2 Thermocouple DTS fiber coating polyimide copper acrylate Plot 3 Thermocouple DTS fiber coating polyimide copper acrylate a

Meter mark along fiber-optic cable where maximum temperature was measured by DTS and thermocouple; Maximum temperature as measured along entire length of fire cable by DTS; c Location of maximum temperature as measured along entire length of fire cable by DTS. b

Sensors 2016, 16,16, 1712 Sensors 2016, 1712

of 12 9 of9 13

In addition, fire cable temperatures across all three fibers on all three burn plots remained In addition, cable temperatures all three on all as three burn plots remained greater greater throughfire time following head, across back and flankfibers fire events compared to thermocouples through time following head, back and flank fire events as compared to thermocouples (Figure (Figure 6). Optical fibers measured greater total thermal energy over time as compared to 6). Optical fibers measured total thermal energy over the timecopper as compared to thermocouples head thermocouples at headgreater fire locations (i.e., 55 m) with fiber generally measuringatthe firegreatest locations (i.e., 55 m) with the copper generally measuring energy (Figure total energy (Figure 7, Table S1).fiber At back fire locations (i.e., the 100 greatest m), total total thermal energy was 7, similar between optical fibers(i.e., and100 thermocouples, withenergy the copper coated fiber still optical measuring theand Table S1). At back fire locations m), total thermal was similar between fibers greatest TTE in most Table Total thermal energyTTE wasinvariable at flank fire 7, thermocouples, with theplaces copper(Figure coated7,fiber stillS1). measuring the greatest most places (Figure locations (i.e.,thermal 123 m) as compared to head at and backfire firelocations locations(i.e., (Figure 7, Table S1). As noted Table S1). Total energy was variable flank 123 m) as compared to head above, thermal mass of theS1). fireAs cable, including strength helps and back the fire greater locations (Figure 7, Table noted above, the the central greatersteel thermal massmember, of the fire cable, explain this response. including the central steel strength member, helps explain this response.

Figure 7. 7. Total thermal and optical opticalfiber fibertemperatures temperatures(°C) (◦ C) from Figure Total thermalenergy energymeasured measuredfor forthermocouple thermocouple and from (a)(a) Plot 1; 1; (b) Plot 33atatdistances distancesalong along cable of(head 55 (head 100 (back fire)123 and Plot (b)Plot Plot2;2;and and (c) (c) Plot thethe cable of 55 fire),fire), 100 (back fire) and 123(flank (flankfire) fire)m.m.

It should be noted that DTS technology, similar to thermocouples, pyrometers, and It should be noted that DTS technology, similar to thermocouples, pyrometers, and calorimeters, calorimeters, cannot record actual flame temperatures, but rather measure its own temperature cannot record actual flame temperatures, but rather measure its own temperature change during change during a fire [3,4,16]. Without extensive knowledge of the thermodynamic properties of the a fire [3,4,16]. Without extensive knowledge of the thermodynamic properties of the fire cable and fire cable and a model for heat transport [17], these fire cable temperatures are not a surrogate for fire a model for heat transport [17], these fire cable temperatures are not a surrogate for fire behavior behavior parameters such as energy density or flux density [2]. When measuring temperatures in parameters energy density density When measuring temperatures in leads turbulent turbulent such flows,as time resolution canorbeflux an issue due[2]. to sub-scale mixing and movement that to flows, time resolution can be an issue due to sub-scale mixing and movement that leads to dramatic fluctuations. If the thermal mass of the fire cable is too great or the time responsedramatic is too fluctuations. the thermal mass of the fire are cable is too great or and the time too must slow these slow these Ifaforementioned fluctuations smoothed over the response resulting isdata be aforementioned are smoothed over and theis resulting data considered an average considered an fluctuations average temperature. However, there still a need to must recordbethese temperatures in temperature. However, theremust is still need to record these temperatures the field [15,18,19] but the field [15,18,19] but they beaquantified as either peak or averaged in temperatures. Bova and they must be quantified as either peak or averaged temperatures. Bova and Dickinson [4] suggested Dickinson [4] suggested temperature measuring instruments like thermocouples, and thereby fire temperature measuring instruments likeintensity thermocouples, and thereby fire cable, can be usedwith to estimate cable, can be used to estimate fireline if calibration is performed by comparison field measurements. fireline intensity if calibration is performed by comparison with field measurements. Potential uses and technology DTS technology fire science continuous spatial Potential uses for for fire fire cablecable and DTS in fire in science includeinclude continuous spatial recording recording of rateduration, of spread, duration, and fire location Prior to this technology, the of rate of spread, and fire location (Figure 8). (Figure Prior to8).this technology, the ability toability capture capture theofdistribution of theseacross parameters across time andrelatively space were relativelyThermocouple impractical. thetodistribution these parameters time and space were impractical. Thermocouple technology for spatial mapping of these fire parameters well, but at technology would allow for would spatialallow mapping of these fire parameters as well, but at as notably limited

scales (i.e., point sampling) and requiring cumbersome wiring. The ability to map and quantify rate of spread and duration may be useful in fire modeling and effects research, respectively. Additional

Sensors 2016, 16, 1712

10 of 13

Sensors 2016, 16, 1712

10 of 12

notably limited scales (i.e., point sampling) and requiring cumbersome wiring. The ability to map and quantify rate of spread and duration may be useful in fire modeling and effects research, laboratory and field experimentation to gain a better are understanding the aefficacy respectively. Additional laboratory are andrequired field experimentation required to on gain better of using fire cable and DTSefficacy technology to quantify parameters. understanding on the of using fire cablethese and DTS technology to quantify these parameters.

(a) Plot 1

(b) Plot 2

(c) Plot 3 Figure 8. Graph illustrating duration of heat for copper coated optical fiber in (a) Plot 1; (b) Plot 2; Figure 8. Graph illustrating duration of heat for copper coated optical fiber in (a) Plot 1; (b) Plot 2; and and (c) Plot 3. Rate of spread can also be illustrated, however, orientation of cable needs to be (c) Plot 3. Rate of spread can also be illustrated, however, orientation of cable needs to be perpendicular, perpendicular, and in this trial the fire cable was parallel to fire spread. All temperatures above 200 and in this trial the fire cable was parallel to fire spread. All temperatures above 200 ◦ C are shaded at °C are shaded at the scale maximum; refer to Figure 5 for maximum recorded temperatures along the scale maximum; refer to Figure 5 for maximum recorded temperatures along cable. Calibration cable. Calibration bath locations were as follows: ice baths at 7–15 m and 146–154 m, and ambient bath locations were as follows: ice baths at 7–15 m and 146–154 m, and ambient temperature baths temperature baths at 16–24 m and 137–145 m. Head, back and flank fires were as follows: 30–73 m, at 16–24 m and 137–145 m. Head, back and flank fires were as follows: 30–73 m, 85–112 m, and 85–112 m, and 118–130 m, respectively. Fire cable that was outside the fire line was as follows: 0–29 118–130 m, respectively. Fire cable that was outside the fire line was as follows: 0–29 m, 74–84 m, m, 74–84 m, 113–117, and 131–154 m. 113–117, and 131–154 m.

It is unlikely that a practitioner would construct a fire cable as complex as the one used in this study. fire cable constructed with construct additionalathermal mass for experimentation It is Our unlikely that awas practitioner would fire cable as complex as the one purposes, used in this but future fire cable cable iterations may avoid thisadditional potentiallythermal challenging Future constructions study. Our fire was constructed with masscondition. for experimentation purposes, a firefire cable caniterations take intomay account learned in this study.condition. For example, thermal mass butoffuture cable avoidlessons this potentially challenging Future constructions challenges can be mitigated as follows: (1) discard the strengthening member; (2) include only one of a fire cable can take into account lessons learned in this study. For example, thermal mass challenges optical fiber thus requiring only one armoring tube; and (3) utilize a smaller diameter armoring tube. can be mitigated as follows: (1) discard the strengthening member; (2) include only one optical fiber These options would reducetube; construction costs. aFurther, offer some thoughts onThese the cost of thus requiring only one also armoring and (3) utilize smallerwe diameter armoring tube. options constructing specialty cables. Often theFurther, exteriorwe armoring is thethoughts most expensive from would also reduce construction costs. offer some on the part cost ranging of constructing

specialty cables. Often the exterior armoring is the most expensive part ranging from $5–10/m. Individual fiber coatings can also add significant cost. The easily available acrylate coated fiber adds $0.01–0.10/m, polyimide coating adds $0.3–5/m, and copper coating adds $20–30/m. Although

Sensors 2016, 16, 1712

11 of 12

the copper coated fiber may outlast the others, the additional cost may not be worth the difference depending on application objectives. 4. Conclusions This study adds to the understanding of using DTS technology for fire science research and establishes a secondary starting point for testing the next generation fire cable at larger spatial scales. The capabilities of this technology may be useful for fire behavior scientists and fire ecologists (Table 3). For example, quantifying fire behavior in heterogeneous fuel beds is difficult, but may be facilitated using this technology. Fiber optic cables act as a single continuous monitoring device that enables representation of fire distribution through spatial domains, something that was practically impossible to achieve with other fire-resistant instruments. Results from this study indicate DTS technology coupled with a fiber-optic fire cable have potential to map and quantify fire parameters such as duration, rate of spread, and fire location at detailed time and space intervals. However, additional laboratory and field testing is required before this technique can be evaluated with regard to these applications. Table 3. Potential strengths and weaknesses of using fire cable and DTS technology to quantify wildland fire behavior parameters, and a summary of alternative methods. Fire Behavior Parameter

Rate of spread (m/min)

Heat duration (time)

Definition

Fire Cable and DTS Technology

Alternative Methods

The linear rate of advance of a fire front in the direction perpendicular to the fire front

Potential Strengths: precise and continuous quantification of time interval between flaming front passage at two points. Weakness: while the fire cable allows for relatively prodigious spatial estimates compared to thermocouples, there are still spatial limitations as fire burns in three dimensions

Hand-held stop watch method is inexpensive but does not produce precision estimates necessary for physics based models. A series of thermocouples, while generally inexpensive, inherently result in spatial gaps that reduce temporal precision and requires numerous wires. Airborne remote sensing lacks the necessary temporal and spatial precision due to resolution limitations

The length of time that heat occurs at a given point. Time at lethal heat (e.g., >60 ◦ C) is often cited

Potential Strengths: given appropriate temperature calibration, greater temporal and spatial quantification as opposed to thermocouples Weakness: similar to thermocouples, temperature calibration is necessary; potential for heat capacity challenges; and cannot easily measure elevated temperatures (e.g., 10 cm above the soil surface)

Thermocouples are frequently used to measure surface heat duration with modest success

Supplementary Materials: The following are available online at http://www.mdpi.com/1424-8220/16/10/ 1712/s1, Figures S1–S9: DTS data for plots 1–3 for each fiber (a) average temperature; (b) standard deviation of temperatures; (c) total thermal energy; and (d) maximum temperature vs. distance along fire cable; Figures S10–S12: DTS data from plots 1–3 copper fiber (a) maximum temperature (◦ C) recorded every 50.6 cm along fire cable; (b) Google Earth image showing location of burn plot; and (c) maximum temperature (◦ C) vs. approximate location; and Table S1: Total thermal energy (TTE) summed during burn at specific location along fire cable. Any additional data not covered by numerical information provided in this paper will be archived at the University of Massachusetts Amherst and may be requested from Christine E. Hatch ([email protected]). Acknowledgments: Thanks to Shad Cox and Richard Dunlap from the Corona Range and Livestock Research Center for access to the study plots and assistance with prescribed fire. Thanks to Brian Herbst formerly of AFL Telecommunications for cable construction and consultation. Thanks to John Selker and the Centers for Transformative Monitoring Programs (CTEMPs) for use of DTS unit. Thanks to the Agricultural Research Station at New Mexico State University for support and funding.

Sensors 2016, 16, 1712

12 of 12

Author Contributions: Scott Tyler, Christine Hatch, Carlos Ochoa and Douglas Cram conceived and designed the experiments; Carlos Ochoa and Douglas Cram performed the field experiments; Christine Hatch, Carlos Ochoa, and Douglas Cram analyzed the data; Scott Tyler contributed the fire cable; Douglas Cram, Christine Hatch, Carlos Ochoa, and Scott Tyler wrote the paper. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3.

4. 5.

6.

7.

8. 9. 10. 11. 12. 13.

14. 15. 16. 17. 18. 19.

Finney, M.A. Calculation of fire spread rates across random landscapes. J. Int. Assoc. Wildland Fire 2003, 12, 167–174. [CrossRef] Kremens, R.L.; Smith, A.M.S.; Dickinson, M.B. Fire metrology: Current and future directions in physics-based methods. Fire Ecol. 2010, 6, 13–35. [CrossRef] Iverson, L.R.; Yaussy, D.A.; Rebbeck, J.; Hutchinson, T.F.; Long, R.P.; Prasad, A.M. A comparison of thermocouples and temperature paints to monitor spatial and temporal characteristics of landscape-scale prescribed fires. J. Int. Assoc. Wildland Fire 2004, 13, 311–322. [CrossRef] Bova, A.S.; Dickinson, M.B. Beyond “fire temperatures”: Calibrating thermocouple probes and modeling their response to surface fires in hardwood fuels. Can. J. For. Res. 2008, 38, 1008–1020. [CrossRef] Selker, J.S.; Thévenaz, L.; Huwald, H.; Mallet, A.; Luxemburg, W.; Van de Geisen, N.; Stejskal, M.; Zeman, J.; Westoff, M.; Parlange, M.B. Distributed fiber-optic temperature sensing for hydrologic systems. Water Resour. Res. 2006, 42, 1–8. [CrossRef] Thomas, C.K.; Kennedy, A.M.; Selker, J.S.; Moretti, A.; Schroth, M.H.; Smoot, A.R.; Tufillaro, N.B.; Zeeman, M.J. High-resolution fibre-optic temperature sensing: A new tool to study the two-dimensional structure of atmospheric surface layer flow. Boundary Layer Meteorol. 2012, 142, 177–192. [CrossRef] Tyler, S.W.; Holland, D.M.; Zagorodnov, V.; Stern, A.A.; Sladek, C.; Kobs, S.; White, S.; Suárez, F.; Bryenton, J. Using distributed temperature sensors to monitor an Antarctic ice shelf and sub-ice-shelf cavity. J. Glaciol. 2013, 59, 583–591. [CrossRef] Sun, M.; Tang, Y.; Yang, S.; Li, J.; Sigrist, M.W.; Dong, F. Fire source localization based on distributed temperature sensing by a dual-line optical fiber system. Sensors 2016, 16, 829. [CrossRef] [PubMed] Liu, Z.; Kim, A.K. Review of recent developments in fire detection technologies. J. Fire Prot. Eng. 2003, 13, 129–151. [CrossRef] Noguchi, K.; Shibata, N.; Uesugi, N.; Negishi, Y. Loss increase for optical fibers exposed to hydrogen atmosphere. J. Lightwave Technol. 1985, 3, 236–243. [CrossRef] Lemaire, P.J. Reliability of optical fibers exposed to hydrogen: Prediction of long-term loss increases. Opt. Eng. 1991, 30, 780–789. [CrossRef] Reinsch, T.; Henninges, J. Temperature-dependent characterization of optical fibres for distributed temperature sensing in hot geothermal wells. Meas. Sci. Technol. 2010, 21, 094022. [CrossRef] Torell, L.A.; McDaniel, K.C.; Cox, S.; Majumdar, S. Eighteen Years (1990–2007) of Climatological Data on NMSU’s Corona Range and Livestock Research Center. Available online: Http://aces.nmsu.edu/pubs/ research/weather_climate/RR761/welcome.html (accessed on 10 October 2016). McDaniel, K.; Hart, C.; Carroll, D. Broom snakeweed control with fire on New Mexico blue grama rangeland. J. Range Manag. 1997, 50, 652–659. [CrossRef] Tyler, S.W.; Selker, J.S.; Hausner, M.B.; Hatch, C.E.; Torgersen, T.; Schladow, G. Environmental Temperature Sensing Using Raman Spectra DTS Fiber-Optic Methods. Water Resour. Res. 2009, 4, 673–679. [CrossRef] Kennard, D.K.; Outcalt, K.W.; Jones, D.; O’Brien, J.J. Comparing techniques for estimating flame temperature of prescribed fires. Fire Ecol. 2005, 1, 75–84. [CrossRef] Neilson, B.T.; Hatch, C.E.; Ban, H.; Tyler, S.W. Effects of solar radiative heating on fiber-optic cables used in aquatic applications. Water Resour. Res. 2010, 46, W08540. Swezy, D.M.; Agee, J.K. Prescribed-fire effects on fine-root and tree mortality in old-growth ponderosa pine. Can. J. For. Res. 1991, 21, 626–634. [CrossRef] Drewa, P.B.; Platt, W.J.; Moser, E.B. Fire effects on resprouting of shrubs in headwaters of southeastern longleaf pine savannas. Ecology 2002, 83, 755–767. [CrossRef] © 2016 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/).