An experimental study on the performance of fire extinguishing system according to the agent charging condition and nozzle type

Abstract

In confined environments such as aircraft, an increase in mass impacts the overall system’s performance, thus requiring sophisticated management. To verify whether the performance characteristics of fire extinguishing systems used in aircraft are satisfied, in this study was built a 1:1 scale test model. We examined the influence of the initial charge state and nozzles. Further, it measured the pressure inside the pipelines and vessels where multiple nozzles are installed to identify the flow and diffusivity characteristics of HFC-125 inside the pipelines and vessels. At a charging ratio of 54%, the initial pressure drop was smaller, and the lowest pressure before the bubble release point appeared 0.26 s later than when the charging ratio was 76%. The average pressure of each nozzle was 275.8 kPa higher under a charging ratio of 54% than 76% and increased further when the average concentration change was 54%, indicating that diffusivity increased. Although improvements occurred according to the charging ratio, the improvements according to the HFC-125 charging mass were more significant.

Graphical abstract

1. Introduction

Various fire extinguishing agents are used depending on the characteristics of a fire. Among the various types of fire extinguishing agents, gas fire suppression systems are used for objects that cannot be exposed to water, such as aircraft or computers in communication stations. Halon 1301 is the primary agent used. Halon 1301 has excellent performance, and its by-products have low toxicity. However, its characteristics as a halogen compound contribute to the depletion of the ozone layer [1]. Consequently, the production of Halon 1301 was banned in most countries under the Montreal Protocol. Thus, it cannot be continuously used except for essential uses [2,3]. Accordingly, the Clean Air Act Amendments (CAAA) and the US Environment Protection Agency (EPA) have restricted the usage of Halon 1301. The Air Force Research Laboratory (AFRL) selected HFC-125 as an alternative based on a comparison of approximately 600 extinguishing agents [4,5].

Nacelle fire is the most difficult to put out. The Federal Aviation Administration (FAA) has studied the spread of fires for aircraft stability [6]. Despite its environmentally friendly performance as an alternative to Halon 1301, HFC-125 requires higher design concentrations of 14.5–26% by volume (depending upon the fire zone configuration), as opposed to 6% for Halon 1301 [7]. This means that a higher quantity of HFC-125 is required when considering the density of each extinguishing agent. The pipe diameter and vessel volume must also be increased according to the weight. In confined environments such as aircraft, an increase in mass impacts the overall system’s performance, thus requiring sophisticated management.

In airplane fire extinguishing systems, 90% of extinguishing agents should be discharged within a very short time for optimal fire extinguishment. In general, for effective fire extinguishment, the discharging velocity of fire extinguishing agents must be increased by using large-sized vessels in which the fire extinguishing agents are highly pressurized by non-combustible gases [8].

As demonstrated in prior research, the uniformity of the concentration of the extinguishing agent is crucial for identifying the extinguishing agent’s performance. It must be maintained for at least 0.5 s above the extinguishable concentration to prevent re-ignition [7]. Bennett et al. (1997) presented extinguishant concentration formulas for the engine nacelle and Auxiliary Power Unit (APU) [7].

The injected agent is affected by the air flowing into the engine nacelle and must have at least the extinguishant concentration in all conditions of supplied air [7,9]. The extinguishant concentration and maintenance time must be satisfied in all sections inside the nacelle.

The agent weight is restricted due to the aircraft's weight limits. And under high-speed ventilation, the air mixed with the agent is quickly discharged outside the aircraft. As such, it is necessary to quickly satisfy the extinguishing performance in all sections with minimal agent weight. Increasing the diffusivity and distribution of the agent injected with each nozzle is critical. For this purpose, it is necessary to understand the state changes of HFC-125 in the vessel and pipelines.

Kim et al. (2009) explained the discharge characteristics of the agent using the Computational Fluid Dynamics (CFD) model corresponding to a case where the rupture disk and pipe diameter were changed in a fire extinguishing system using Halon 1301. Amatriain et al. (2021) calculated the injection speed of the agent using the charging mass and charging temperature of Halon 1301 as variables [10].

The interior of an aircraft nacelle is highly complex, and complex flow occurs due to a large amount of intake air. Niu et al. (2013) presented a fire spread scenario using CFD in a helicopter nacelle using Halon 1301 under no-ventilation conditions [6]. By comparing agent concentration measurements, Lee (2004) investigated simulation methods for predicting concentration in the aircraft nacelle using the injection characteristics of the fire extinguishing agent as variables [11]. Using CFD, Lopez et al. (1997) presented the global flow pattern and mixing of air and the suppression agent according to the agent’s injection position inside the nacelle of the aircraft, excluding internal components [12].

The capacity of the fire extinguishing system installed on aircraft must be designed based on the maximum air inflow. Park et al. proposed a prediction model for inflow air by applying the model used in this study [13], and Hamins et al. presented the correlation between inflow air and the minimum agent mass [14]. Cho et al. proposed a prediction model for the concentration of HFC-125 in the nacelle used in this study [15].

Elliott et al. (1984) presented a model on pressure and internal state changes inside a fire extinguishing system's vessel and pipelines using Halon 1301. They compared the theoretical model with experimental results [16].

Jin et al. (2020) confirmed the changes in agent concentration according to nozzle type under no-ventilation conditions using halogenated alkane [17]. To determine whether the extinguishant concentration was satisfied, the FAA and the National Institute of Standards and Technology (NIST) divided the inside of a nacelle into 12 sections and measured the agent concentration [18,19].

The air flowing into the nacelle varies with the aircraft’s flight conditions [20], and it is difficult to conduct the test during flight. Keyser et al. measured the airflow and average pressure of flight conditions and compared and verified the prediction model with a nacelle simulator for ground tests [21].

Oguchi et al. studied the state quantity of HFC-125 [22], and Keyser et al. tested the discharge rate and fire suppression of HFC-125 at various charging pressures. However, they did not consider the influence of dissolved nitrogen [23]. showed that the diffusivity increased by measuring the speed and range of CO2 Jet when changing from a liquid phase to a gas phase [24].

To identify the flow and diffusivity characteristics of HFC-125 inside the pipelines and vessels, this study measured the pressure inside the pipelines and vessels where multiple nozzles were installed. It is also measured the concentration of HFC-125 inside the nacelle under high-speed ventilation. Nozzle type, vessel volume, HFC-125 charging ratio, HFC-125 charging mass, vessel pressure, and nitrogen saturation state are considered variables.

2. Experimental setup and conditions

2.1. Experimental setup

Fig. 1 shows the composition of test apparatus. Fig. 1 (a) shows a schematic diagram of the fire extinguishing test system. The pressure sensors are marked with P, and the concentration measuring points are numbered clockwise. The pressure sensors are numbered along the flow direction of the agent. A total of 13 pressure sensors (all sensors; P201GB) were installed in the vessel, pipelines, and nozzles to monitor the state of these components.

Fig. 1.

Composition of test apparatus: (a) Schematic diagram of test apparatus; (b) position of pressure measuring points and nozzles; (c) composition of nozzle block.

Fig. 1 (b) shows the installation positions of the pressure sensors and nozzles. A total of four nozzles were installed: one in the gear box (GB) (Nozzle 1), two in the forward region (FWD) (nozzle 2, nozzle 3), and one in the middle region (MID) (nozzle 4).

Fig. 1 (c) shows an example of the pipeline and nozzle block configuration, corresponding to pipeline e in Fig. 1 (a). Each nozzle was processed in the form of a block to extend the internal diameter of the pipeline. Further, it was fabricated to enable pressure measurements and nozzle replacement.

Table 1 shows the test conditions (vessel pressure and volume, weight of HFC-125, nozzle’s internal diameter) and the peak pressure (P.) at each nozzle. Table 1 presents the diameter (D.) of each nozzle. In Table 1, straight nozzles were used for test 1, and convergent nozzles for tests 2–10. In Test 1 and Test 2, a straight nozzle and a convergent nozzle were used. The pressure in the pipeline and nozzle and the concentration in the nacelle were measured and compared. Test 3–6 were experiments to compare the effectiveness of charging ratio. In Test 3 and Test 4, 5.9 L and 8.4 L vessel were used with an HFC-125 charging ratio of 54%. Test 5 and Test 6 used 5.9 L containers and 8.4 L containers with a 76% HFC-125 charging ratio. Tests 3, 4 and Tests 5 and 6 each increased the weight of HFC-125 at the same charging ratio. Test 4 and Test 5 had the same HFC-125 weight. Tests 1–6 were tested under nitrogen saturated conditions, and Tests 7–10 were tested under nitrogen unsaturated conditions. The difference between Test 5 and Test 7 was nitrogen saturation. Test 8–10 confirmed the pressure change in the vessel by the charged pressure.

Table 1.

Test condition and Results.

Test No.	Vessel Pressure	Vessel Volume	Nozzle type	Nozzle 1		Nozzle 2		Nozzle 3		Nozzle 4
	Dissolved Nitrogen	HFC-125 Weight	Charge ratio	D. (mm)	P. (MPa)	D. (mm)	P. (MPa)	D. (mm)	P. (MPa)	D. (mm)	P. (MPa)
1	5.965 MPa	5.9 L	SN	7.5	0.883	15.6	0.676	15.6	–	7.5	1.083
	saturated	5.47 kg	76%
2	5.599 MPa	5.9 L	CN	4.3	2.027	7.8	2.027	11	2.069	11	2.351
	saturated	5.47 kg	76%
3	4.233 MPa	5.9 L	CN	4.3	2.033	9.0	1.958	11	2.041	9.0	2.599
	saturated	3.88 kg	54%
4	4.260 MPa	8.4 L	CN	4.3	2.055	9.0	2.041	11	2.172	9.0	2.737
	saturated	5.48 kg	54%
5	4.206 MPa	5.9 L	CN	4.3	1.848	9.0	1.806	11	1.903	9.0	2.344
	saturated	5.46 kg	76%
6	4.213 MPa	8.4 L	CN	4.3	1.937	9.0	1.882	11	1.951	9.0	2.448
	saturated	7.75 kg	76%
7	4.171 MPa	5.9 L	CN	4.3	1.593	9.0	1.607	11	1.627	9.0	1.889
	unsaturated	5.48 kg	76%
8	4.130 MPa	8.4 L	CN	7.5	–	15.4	–	15.4	–	7.5	–
	unsaturated	5.45 kg	54%
9	3.440 MPa	8.4 L	CN	7.5	–	15.4	–	15.4	–	7.5	–
	unsaturated	5.45 kg	54%
10	2.758 MPa	8.4 L	CN	7.5	–	15.4	–	15.4	–	7.5	–
	unsaturated	5.45 kg	54%

D.: Nozzle Diameter (mm).

P.: Peak Pressure at Nozzle (Point 6; MPa).

SN: Straight Nozzle.

CN: Convergent Nozzle.

HFC-125 begins charging with the agent stopper in a closed state. Table 1 shows the charging mass of HFC-125 for each test. A scale (AND; 150 KAL-H) was used for the agent charging to measure the weight of the vessel before and after charging. After charging the agent, the inside of the vessel was pressurized with nitrogen. And The vessel was repeatedly shaken until the nitrogen was no longer dissolved. Then, nitrogen saturated condition was confirmed that the pressure was maintained during 2 h. It is to confirm the pressure recovery phenomenon by dissolved nitrogen. The final charged pressure is marked as “vessel pressure” in Table 1.

Two types of vessels were used with volumes of 5.9 and 8.4 L. Table 1 shows the type of vessel used for each test. The agent stopper installed at the vessel outlet was designed such that the extinguishing agent could be instantaneously released. This study used a pneumatic valve (13.790 MPa) for the agent stopper, and the opening and closing times of the valve were 0.15 s.

When the agent stopper is opened, the HFC-125 inside the vessel flows along the pipeline and is injected into the test bed through the nozzle. The injected HFC-125 diffuses, mixes with the inflow air of Fig. 1 (a), and is discharged through Outflow 1 and 2. The flow rate of the inflow air is 0.96 kg/s, and the agent is discharged through Outflow 1 at 0.64 kg/s and Outflow 2 at 0.32 kg/s.

Due to the high flow rate of the inflow air, a turbo blower (TURBOWIN; WL100, 82 kW) was used to supply air. To precisely control the airflow, 200 A butterfly valves were installed at the inlet and outlet, and 100 A bypass valves were additionally installed. Straight pipes of at least 3 m were installed at the air supply inlet, and outlet for the flow to sufficiently develop, and a mass flow meter (Sierra; 640i) was installed to measure the flow rate. The maximum error of the supply air between tests is $\pm$ 2%. To verify the similarity with supply air under the flight condition, we measured the total pressure for inflow and outflow using a pressure scanner (SCANNIVAVLE; DSA3217) and pitot tube (KIMO AFNOR).

The test bed simulated the aircraft's nacelle at a 1:1 scale. Major components were simply manufactured and arranged to increase similarity, but the wiring was not simulated.

Agents were uniformly arranged so that concentration uniformity in each section could be measured, as in Fig. 1 (a) [18,19]. For the position selection method, the nacelle was divided into three equal sections FWD, MID, and AFT, each section was divided into four equal segments, and the center of each segment was measured. Considering the airflow swirl, the MID section's measuring point was determined with reference to Ref. [18]. To account for the extinguishant concentration, the area around the engine core with the fastest air flow velocity was selected. The concentration was measured at the height of 10 mm from the engine core.

2.2. Verification of experimental conditions

In the aircraft simulated in this study, an external flow inlet introduces airflow from the aircraft surface into the nacelle for ventilation. The amount of external air introduced through this inlet varies with the flight conditions [20]. Keyser et al. measured the airflow and average pressure of flight conditions and compared and verified the prediction model with a nacelle simulator for ground tests [21].

Table 2 compares the predicted model and test model pressure with airflow conditions according to the flight condition. To determine the upper limit of the HFC-125 charging mass, we selected the boundary condition with the most airflow under each flight condition. The airflow of the predicted model was derived through a numerical study by modeling a test bed [13]. As the mass flow of air increases, the minimum agent mass increases linearly [14]. According to the numerical study results on concentration with the predicted model, as the airflow increased, the concentration maintenance time decreased, thus increasing the required charging mass of HFC-125 [15].

Table 2.

Comparison of Pt ratio (predicted and test).

Mass flow rate (kg/s)			Pt Ratio		Error (%)
Inflow	Outflow 1	Outflow 2	Predicted Result	Test Result
0.46	−0.18	−0.27	0.98	0.98	0
0.68	−0.23	−0.45	0.96	0.96	0
0.83	−0.36	−0.47	0.94	0.95	0.553
0.82	−0.40	−0.42	0.94	0.95	1.030
0.84	−0.49	−0.34	0.93	0.95	2.105
0.96	−0.64	−0.42	0.91	0.94	3.192

The airflow according to flight conditions was supplied to the test model, and the mass flow rate was measured. The temperature and total pressure of the air inlet and outlet were also measured to compare the total pressure ratio of inflow and outflow. Table 2 shows the total pressure measurements of the predicted and test models.

The Pt ratio of Equation (1) is the average total pressure ratio of Inflow and Outflow 2. The Pt ratio error of Equation (2) is the difference between the predicted and test models. Three pitot tubes were installed in Inflow and Outflow 2 each to measure the total pressure of the test model, after which the average was taken. The flow rate was measured under 0.96 kg/s; the maximum Pt ratio error was 3.1%. The result is judged to be an error caused by flow rate control of inflow and outflow. It was lower than previous study (4.5%). It means that the predicted model and test model Pt ratio with airflow is similar. Regarding air mass flow rate, the conditions inside the test bed are similar to the inside of the predicted model.

$$Ptratio=\frac{P_{t,avg.ofairoutflow2}}{P_{t,avg.ofairinflow}}$$ (1)

$$Ptratioerror=\left|\frac{Ptratio,testmodel-Ptratio,predictedmodel}{Ptratio,testmodel}\right|\times 100(\%)$$ (2)

Bennet et al. presented the extinguishant concentration formula shown in (2) and revealed that certification design concentration (Xe) must be satisfied for at least 0.5 s in all measured sections [7]. Xe, AIRT, $\dot{W}$ a, FUEL CONSTANT mean certification design concentration, maximum ventilation air temperature, internal air mass flow rate in the nacelle and coefficient to account for the presence of oil. Xe is determined according to the mass flow rate, for which 0.96 kg/s was used in this study. Xe derived from Equation (2) is 18.8% in the nacelle and 22.6% in the GB. To satisfy the fire suppressing performance, the concentration of Xe must be measured for at least 0.5 s at all concentration measuring points.

$$Xe=21.10+0.0185AIRT-3.124\dot{W}a+5.174\left(FUELCONSTANT\right)$$

$$+0.0023\left(AIRT\right)\times \left(FUELCONSTANT\right)+{1.597\left(FUELCONSTANT\right)}^2$$ (3)

3. Results and discussion

3.1. Effects of nozzle type on HFC-125 concentration

Fig. 2 shows the pressure measurements after injecting HFC-125 with a straight and convergent nozzle. Eliott and Jin et al. explained the state change inside the vessel and pipeline using Halon 1301, a halogenated alkane [16,17]. HFC-125 used in this test also showed a similar state change. Fig. 2 (a) shows the experimental results of applying the straight nozzle. At point 1, the stopper is opened, and HFC-125 begins to flow from the vessel. The pressure in P1 rapidly drops, and the pressure in the pipe and nozzle rises. Inflection point 5 is where liquid HFC-125 reaches each pressure measuring point, and point 6 is the peak pressure at each measuring point of the pipe and nozzle. Liquid HFC-125 forms the highest pressure at pressure sensor 3 installed against the flow direction. It then branches out to b and e in Fig. 1, and the pressures at P5 and P12 similarly form. The HFC-125 flowing along b does not sequentially form a pressure drop and forms the lowest pressure (204.1 kPa) at P6. The pressure at the measuring point located after P9 increases due to the 77% reduction in the cross-sectional area of Nozzle 1. At P6, the inflection point (point 5) where liquid HFC-125 reaches is very unclear, and the pressure is 204.1 kPa, which is markedly lower than the vapor pressure of HFC-125 (1.20 MPa at 20.0 °C) [22]. Consequently, after pipe b, it is vaporized and discharged in a two-phase state. The diffusivity of HFC-125, when injected from the nozzle in a vapor state, is substantially lower than when injected in a liquid state. This is because the injection pressure at the nozzle is basically lower, and expansion due to vaporization affects the diffusion rate. The peak pressure of P7 to P9 decreased linearly, and the pressure of Nozzle 2 significantly decreased to 31% of that of Nozzle 3. This means that the mass flow rate of Nozzle 2 significantly decreases compared to Nozzle 3. Table 3 shows summary of major test devise. Manufacturer, specification, accuracy, purpose of use, and quantity were indicated. The maximum error is 2% at the blower, which is due to fluctuation in the characteristics of the blower. VF(Volume Fraction) was measured using a sensor from Pyroscience. The concentration was measured with OXR-430CV and the temperature was measured with TPR-430CV for correction. An explanation was added as shown in Table 3.

Fig. 2.

Pressure record by all sensors: (a) Used straight nozzle test 1; (b) Used convergent nozzle test 2.

Table 3.

Summary of major test devices.

Name	Quantity	Purpose	Manufacturer/specification	Error
Blower	1ea	Supply airflow inlet	TURBOWIN; WL100 82 kW	$\pm$ 2%
Mass flow meter	3ea	Measure airflow	Sierra; 640i 305smps	$\pm$ 0.75%
Concentration sensor	16ea	Measure agent concentration	Pyroscience; OXR-430CV 0–100%	±0.2%
Temperature sensor	16ea	Measure agent concentration compensation	Pyroscience; TPR-430CV 0–50 °C	±1%
Scale	1ea	Measure agent weight	AND; FC-200KI 0–150 kg	Under ±0.6%
Pressure sensor	13ea	Measure agent pressure	All Sensors; P201GB 0–1000 psig	±0.25%
Pressure scanner	1ea	Measure airflow pressure	SCANNIVAVLE; DSA3217	±0.05%
Pitot tube			KIMO AFNOR	–
Pneumatic valve	1ea	Agent stopper	BD52/HS3, Valman valve 2,000psig	–
DAQ1	1ea	Gether Pressure sensor data	NI; 9205 sampling rate: 15,625 Hz/Ch.	–
DAQ2	2ea	Gether Concentration sensor data	Pyroscience; Fire-Sting Sampling rate: 10 Hz	–
Computer	1ea	Record data	MSI; I7	–

Fig. 2 (b) shows the results of the experiment applying the convergent nozzle. The pressure at the lowest point P6 is 1.724 MPa, which is higher than the vapor pressure (1.20 MPa at 20.0 °C). Oguchi et al. experimented pressure-volume-temperature properties and vapor pressures of HFC-125 [22]. Table 4 shows the vapor pressure at various temperatures of HFC-125 [22]. Hence, most of the HFC-125 is injected in the liquid phase from all nozzles. The peak pressure is the lowest at P6 due to a complex phenomenon in which some HFC-125 vaporizes at the changing edge of the pipe cross-section, and the equilibrium bubble generated during pressure recovery escapes to the vertical Pipe b just before P6 in the direction of vessel. In a filled state, the pressurized liquid in the pipe branches from the branch point of P3 to pipes on both sides; P4 and P12 form the same pressure, as well as P5 and P13. In addition, P7, P8, P9, P10, and P11 form similar pressures. In this case, the mass flow rate of the agent in Nozzles 1, 2, and 3 is proportional to the cross-sectional area of each nozzle, and the HFC-125 concentration distribution inside the nacelle can be controlled using the diameter of the nozzle. The mass flow rate of HFC-125, which is injected into the branching pipes and MID section, can also be adjusted concerning other pipes and nozzles. Particularly, since the influence of airflow is negligible in the GB, the concentration of the GB section is determined only by the mass flow rate of HFC-125 injected through Nozzle 1. When the mass to the GB section is minimized, HFC-125 injected into the nacelle is increased; thus, the concentration distribution of the nacelle can be improved, and the agent charging mass can be optimized. Optimizing the charging mass of HFC-125 and piping has the positive effect of reducing the total system weight.

Table 4.

Vapor pressure for HFC-125 [22].

Temperature (K)	Pressure (MPa)
223.160	0.0929
233.159	0.1486
233.159	0.1487
243.146	0.2285
243.156	0.2283
253.151	0.3380
253.152	0.3367
253.154	0.3378
253.154	0.3379
263.147	0.4832
263.149	0.4830
263.152	0.4830
263.152	0.4833
273.146	0.6710
273.150	0.6710
273.150	0.6712
273.151	0.6714
283.144	0.9088
283.144	0.9092
293.141	1.2047
293.141	1.2056
293.184	1.2066
298.144	1.3774
298.183	1.3800
303.139	1.5678
303.139	1.5679
303.139	1.5692
303.181	1.5701
303.181	1.5705
313.136	2.0075
313.136	2.0077
313.136	2.0078
313.136	2.0080
313.136	2.0082
313.136	2.0089
313.136	2.0090
313.136	2.0094
317.148	2.2090
318.138	2.2606
319.638	2.3411
323.133	2.5360
323.133	2.5363

Fig. 3 shows the average concentration change of each section measured using the straight and convergent nozzles. Straight nozzles were used for test 1, and convergent nozzles for test 2. When convergent nozzles were used, the measured concentration of HFC-125 was higher overall, indicating improved diffusivity. The peak concentrations of FWD and MID appeared faster than in test 1. This is because, in Table 1, the injection pressure of nozzles 2, 3, and 4 increased, so it was injected as a liquid and vaporized inside the nacelle, thus increasing the diffusion rate. As there are no separate nozzles in the AFT section, the upstream HFC-125 diffuses with the airflow to AFT, so it is influenced only by the airflow velocity and dispersion by the injection angle of nozzle 4. As a result, the concentration of AFT increases after a lag. The influence of airflow is negligible in the GB, and when HFC-125 is injected, the same concentration is maintained for an extended period. Therefore, the concentration of the GB is proportional to the mass of HFC-125 injected through Nozzle 1. In Fig. 3, when using straight nozzles, the HFC-125 peak concentration of the GB is 41.1%, and the mass injected from Nozzle 1 is excessively high. In contrast, when using convergent nozzles, the distribution from Nozzle 1 is about half at 20.4%. Thus, the concentration of the injection amount does not reach the target concentration of 22.6%, so the injection amount to GB must be increased.

Fig. 3.

Comparisons of HFC-125 average concentrations in system with different nozzles.

3.2. Effect of charging ratio and dissolved nitrogen

Fig. 4 shows the results of an experiment in which 5.9 L and 8.4 L vessels were charged with HFC-125 to 54% and 76%, respectively. Fig. 4 (a) shows the result of comparing the pressure from 0 to 6 s. Fig. 4 (b)–(e) shows enlarged graphs of 0–2.5 s, 1.5–3 s, 3–6 s, and 2.9–3.3 s, respectively. Table 1 shows the HFC-125 charge mass and vessel for each experiment. In tests 3–6, the vessel was shaken until nitrogen was no longer dissolved, and HFC-125 was charged at 54% and 76% rates. In tests 7–10, for a comparison with previous studies, the vessel was not shaken, and only natural dissolution occurred on the surface [23]. At inflection point 2 in Fig. 4, the lowest pressure occurs before pressure recovery due to bubble release. Inflection Point 3 is the equilibrium bubble pressure reached the point, and at point 4, the discharge of liquid HFC-125 ends. As shown in Fig. 2, point 4 is the inflection point of the pressure measured at the nozzle.

Fig. 4.

Comparison of pressure at vessel: (a) Time 0–6sec; (b) Time 0–2.5sec; (c) Time 1.5–3sec; (d) Time 3–6sec.

The time to reach (x-axis) inflection points 2 and 3 differed with the charging ratio of HFC-125. When charged to 54% (red line; test 3, 4), point 2 appeared at 0.46 s after the valve was opened, and point 3 appeared at 0.57 s. When charged to 76% (blue line; test 5, 6), point 2 appeared at 0.20 s, and point 3 appeared at 0.30 s. These results show no difference in the time to reach inflection points 2 and 3 under identical charging ratios. The difference in time to reach the inflection point for each charging ratio is attributed to the difference in the shape of the 5.9 L container and 8.4 L vessel. The same pipelines and nozzles were used in the experiments. For the 5.9 L vessel, the agent charge port is inserted into the spherical vessel. There are also three agent discharge ports, two of which were closed. For the 8.4 L vessel, a charge port on the outside of the vessel and an agent discharge port are present. The internal volume ratio between nitrogen and the agent in each vessel is the same. In tests 4 and 5, the HFC-125 charging mass is identical, but the charging ratios differ. In test 4 (54% agent charging ratio), the time to reach Point 4, where the liquid injection ends at the nozzle, is 1.27 s, and in test 5 (76% agent charging ratio), the time to reach Point 4 is 1.43 s, showing that liquid injection ends more quickly. This is because, as the charging ratio of HFC-125 decreases, the ratio of nitrogen used as a pressurizing agent increases, causing a higher pressure to be maintained in the vessel. However, the time difference at point 4 between the two tests was insignificant, and the time the inside of the vessel took to be emptied was the same. Since the pipelines and nozzles are the same, the mass flow rate of HFC-125 is limited. This point can be confirmed from the measurement results of pressure changes inside the vessel in tests 4, 5, 7, and 8 (Fig. 4; (e) mark). In these tests, the points where HFC-125 was charged under the conditions of approximately 4.137 MPa (600 psig) and 5.48 kg were identical, but the vessel, charging ratio, and charged method were different. However, when including the remaining HFC-125 in a vapor state, the time from the valve opening to releasing of the inside of the vessel was similar at 3.73–3.90 s. This error in the time for the inside of the vessel to release is attributed to the difference in initial pressure of 131.0 kPa. A prior study explained this as the discharge due to unknown phenomena in the vessel and pipe, which is proportional to the pressure [23]. In tests 8, 9, and 10, the charged pressure was changed at 689.5 kPa increments, and the test time was limited to 20 min after charging, through which identical levels of dissolved nitrogen were maintained. 4.45 kg of HFC-125 is discharged at 4.137 MPa (600 psig), 3.488 MPa (500 psig), and 2.758 MPa (400 psig) over 3.9 s, 4.4 s, and 5.0 s, respectively, and the difference in discharge time due to the initial pressure is shown. This can be utilized as primary data for designing the charged pressure of the vessel.

The initial pressure change ΔP_init signifies the internal pressure drop due to nitrogen expansion caused by the discharge mass of HFC-125 from the vessel. When the nitrogen charging ratio is high, the pressure drop inside the vessel according to the discharge mass of HFC-125 from the vessel is small. As the charging ratio of HFC-125 increases (76%, 0.20 s; 54%, 0.46 s), the time to reach the initial lowest pressure before the bubble release point (point 2) decreases, and pressure recovery becomes stronger. ΔP_recovery was 35.2 kPa in test 3, 146.9 kPa in test 4, 278.6 kPa in test 5, and 570.2 kPa in test 6; thus, Test 5 (76% HFC-125 charging ratio) recovered 131.7 kPa more than test 4 (54% HFC-125 charging ratio). This strong pressure recovery occurs because the dissolved nitrogen generates bubbles due to rarefaction, thus increasing the volume of the liquid. This is suppressed by the pressure of nitrogen charged as the pressurizing agent [16,17]. 3.88 kg of HFC-125 is charged in test 3, 5.48 kg in tests 4 and 5, and 7.75 kg in test 6. Test 4 showed 111.7 kPa more pressure recovery than test 3, and test 6 showed 291.7 kPa more than test 5. This is because, in these tests, the injection pressure at the nozzle was higher than the vapor pressure, thus limiting the discharge rate of HFC-125.

Pressure recovery increases the pressure in the pipe, enabling injection at a pressure higher than the vapor pressure from the nozzle. An injection pressure higher than the vapor pressure fills the inside of the pipeline with liquid and makes the injection pressure between the nozzles uniform. During liquid injection, the diffusivity of HFC-125 injected from each nozzle increases. Test 7 is performed under the same conditions as test 5 without fully dissolving the nitrogen. The equilibrium bubble caused by dissolved nitrogen was limited, and a clear inflection point due to pressure recovery did not appear. The pressure at infection point 3 of test 5 was 3.126 MPa, whereas the pressure at P1 of test 7 at the same time was 71.1% of test 5 at 2.224 MPa, indicating that pressure did not recover and a sharp pressure drop occurred. Table 1 shows the peak pressure (point 6) measured at each nozzle. The peak pressure of each nozzle varied with the charging ratio. The average peak pressures at the nozzle in tests 3, 4, 5, and 6 were 2.158 MPa, 2.251 MPa, 1.975 MPa, and 2.055 MPa; thus, the peak pressures of tests 3 and 4 (54% HFC-125 charging ratio) were higher. This is because the pressure reduction of nitrogen, the pressurizing agent in the vessel, is small. Moreover, for the same reason, the pressure at the nozzle increases as the vessel volume increases with identical HFC-125 mass. This is because there is no change in pipe diameter, and the discharge rate of HFC-125 is limited due to choking, etc. As the capacity of the pressurizing agent (nitrogen) increases, the pressure drop in the vessel, the injection power source of the agent, decreases. A comparison of tests 7 and 8 shows a more striking difference. Though the cross-sectional area of the test 7 nozzle is 51.5% smaller than test 8, the pressure of P1 is higher in all sections of test 8.

Fig. 5 shows the average concentration change in each section of tests 4 and 5, which have the same HFC-125 charging mass but different charging ratios. Test 4 (54% HFC-125 charging ratio) showed a slightly higher initial concentration than test 5 (76% charging ratio). This is attributed to the difference in injection pressure at the nozzle. In Table 1, the peak pressure of the test 4 nozzle is 206.9–393.0 kPa higher than that of test 5. In Fig. 5, a higher vessel pressure is maintained in test 4 than in test 5. This pressure difference causes a slight difference in the concentration peak and maintenance time due to diffusivity, although the overall trends are similar.

Fig. 5.

Comparisons of HFC-125 average concentrations in system with different HFC-125 charging ratio: Test 4 54%; Test 5; 76%.

In the previous test results, when HFC-125 was charged to 54%, the pressure in the vessel, pressure maintenance in the nozzle, diffusivity, and concentration results was slightly better, although the volume of the vessel increased. Given that the weight that can be loaded on aircraft is limited, this has a more significant disadvantage of degrading mounting performance and increasing weight. The difference caused by the charging ratio was insignificant, particularly regarding concentration, the ultimate target, whereas HFC-125 charging mass showed a more significant influence. The injection pressure drop at the nozzle due to the charging ratio was also small at 206.9–393.0 kPa, and according to the comparison of tests 8, 9, and 10, it is possible to compensate for the pressure drop at the nozzle by increasing the charged pressure (increasing the nitrogen mass).

Fig. 6 shows the comparison of the concentration of hfc-125 according to charging mass of HFC-125. Fig. 6 (a) shows a comparison of the average concentration change between test 4 (8.4 L vessel charged with 5.48 kg of HFC-125) and test 6 (8.4 L vessel charged with 7.75 kg of HFC-125). As shown, the average concentrations in test 6 (76% charging ratio) exceeded those of test 4 (54% charging ratio) in all sections. Considering the difference in concentration of the GB, the HFC-125 concentration level of the nacelle can be further increased by limiting the injection amount through Nozzle 1. Compared to test 4–6, which compares the average concentration change according to charging ratio, demonstrating that the influence of the HFC-125 charging mass is more significant. Fig. 6 (b) shows the concentration changes of all sensors in the nacelle of test 6. Fig. 6 (c)–(e) shows the concentration results in the FWD, MID, and AFT sections. In the extinguishant concentration formula (2), the certification design concentration (Xe) must be satisfied for at least 0.5 s in all measured sections [7]. In test 6, the Xe concentration was 18.8%, and the concentration peak and maintenance time increased in all sections compared to those in test 4, confirming a 0.8-s interval where the concentration of at least 18.8% was satisfied in all sections. This signifies that the fire was suppressed in all sections.

Fig. 6.

Comparisons of HFC-125 concentrations in system with different HFC-125 charging mass: (a) Average concentration: Test 4 5.48 kg; Test 6; 7.75 kg; (b) Concentration with all sensor: Test 6; (c) Concentration of FWD section; (d) Concentration of MID section; (e) Concentration of AFT section.

4. Conclusion

In aircraft, the installation volume and weight of the fire extinguishing system are limited, requiring an optimized design. In this study, the characteristics of the HFC-125 according to charged condition were studied for an optimized design. The 1:1 scale test bed is built to create an environment similar to the actual aircraft, a high-speed ventilation environment was established, and various variables were tested using multi-nozzles.

• 1)Compared to the straight nozzle, the convergent nozzle had an average pressure of 242% higher at the nozzle.
• 2)The minimum pressure in the pipeline of the Convergent nozzle is 1.724 MPa. The pressure was maintained higher than the HFC-125 vapor pressure (1.20 MPa at 20.0 °C) in all nozzles.
• 3)When using the straight nozzle, the maximum pressure in the pipe was 1.083 MPa, which was lower than the vapor pressure of HFC-125 (1.20 MPa at 20.0 °C).
• 4)The HFC-125 charge ratio is involved in the initial Pressure drop, bubble release (point 2), and HFC-125 Liquid discharge time. In the case of HFC-125 charging ratio of 54%, the pressure drop was less, the bubble release appeared later, and the HFC-125 Liquid was discharged more rapidly than in the case of 74%.
• 5)When the HFC-125 charging ratio was 54%, the average pressure of nozzles was 275.8 kPa higher than when the HFC-125 charging ratio was 76%. It was confirmed that the diffusivity increased as the average concentration increased more rapidly.
• 6)Dissolved nitrogen causes pressure recovery in HFC-125.
• 7)The Pressure recovery were stronger in the nitrogen saturated condition. In these case, the pressure at the nozzle increased, hence increasing diffusivity, and the liquid phase of HFC-125 was quickly discharged to increase the initial discharge rate.
• 8)Charged pressure was involved in the time when the Vessel was empty.
• 9)If it is necessary to adjust the concentration distribution in the test volume, it is possible by adjusting the discharge rate. This can be done by adjusting the charged pressure, charging ratio, vessel volume, nozzle, and HFC-125 wt
• 10)Comparing the effect of the charging ratio with the effect of the increase in HFC-125 mass, the increase in HFC-125 mass is more effective in the weight design of the system. Therefore, it is necessary to first find the optimal agent charge weight. After that, the weight of the system must be lowered by adjusting the charging ratio, nozzle type and diameter, and vessel volume.
• 11)Extinguishant concentration was satisfied 0.8 s in Test 6 (charging ratio 76%, HFC-125 charge weight 7.8 kg, charged pressure 4.2 MPa, Vessel volume 8.4 L).

The fire extinguishing system using HFC-125 which is an eco-friendly refrigerant shall be designed and verified as follows;

• a)Calculate maximum airflow inlet (Worst case)
• b)Calculate required HFC-125 wt under maximum airflow inlet condition
• c)Design the agent charged pressure
• d)Design the discharge rate of pipe lines and nozzles
• e)Verify by simulations and experiments
• f)Optimize the system weight (Vessel volume, agent weight, chare ratio etc.)

HFC-125 has been applied recently. There were few examples of experiments and applications. Moreover, there were few cases of experiments with multiple nozzles, high speed ventilation environments, and 1:1 scale. In this study, there was trial and error in step b)-d). In test 6, the extinguishant concentration was satisfied with 0.8s interval. Aircraft's Fire Extinguish system should be optimized because the weight is limited. In addition, the agent weight and discharge rate must be adjusted according to the internal shape and the position of the nozzle. In this study, variables such as vessel volume, nozzles, HFC-125 wt, charging pressure, charging ratio, and nitrogen charging state were tested to identify what occurred during the installing process. This will aid in the design and optimization of the system.

To verify whether the performance characteristics of fire extinguishing systems used in aircraft are satisfied, we built a 1:1-scale test model. We examined the influence of the initial charge state and nozzles. Other factors impacting the performance of fire extinguishing systems include the ventilation flow rate, nozzle, pipeline diameter, and ambient temperature. Although this study was conducted under the worst-case condition with the most significant ventilation flow rate in various flight environments, the change in the airflow of the internal swirl and resulting concentration diffusion characteristics of the agent change. This internal airflow change is due to differences in the internal shape. Moreover, although the currently configured condition uses a valve as the stopper, the flow characteristics inside the pipelines may differ if a ruptured disc is used. A follow-up study will use a ruptured disc to investigate the flow characteristics inside pipelines and various nozzles for controlling these characteristics, in addition to investigating the influence of the pipeline configuration. The internal shape will be changed to confirm the relation with the required agent and develop a prediction model.

Author contribution statement

Yeseul Park: Conceived and designed the experiments.

Cheolhee Ahn, Hyeongjin Ahn & Jiwon Park: Performed the experiments.

Warngyu Park: Analyzed and interpreted the data.

Cheolung Cheong & Gyungmin Choi: Contributed reagents, materials, analysis tools or data.

Junsung Kim: Wrote the paper.

1 - Conceived and designed the experiments.

2 - Performed the experiments.

3 - Analyzed and interpreted the data.

4 - Contributed reagents, materials, analysis tools or data.

5 - Wrote the paper.

Data availability statement

The authors are unable or have chosen not to specify which data has been used.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Nomenclature

CFD

computational fluid dynamics

CAAA

Clean Air Act Amendments

EPA

U.S. Environment Protection Agency

AFRL

Air Force Research Laboratory

F.A.A.

Federal Aviation Administration

NIST

National Institute of Standards and Technology

P.

Peak pressure

D.

Nozzle Diameter

GB

Gear box

FWD

Forward region

MID

Middle region

AFT

After region

Dia.

Diameter

s

Seconds

Pt

Total pressure

M

Mach number

Xe

Certification Design Concentration

AIRT

maximum ventilation air temperature in the nacelle or APU during operations

$\dot{W}$ a

internal air mass flow rate in the nacelle or APU during operations

FUEL CONSTANT

coefficient to account for the presence of JP fuel, hydraulic fluids, or oil

L

liters

P

Pressure

Init

initial

SN

Straight Nozzle

CN

Convergent Nozzle