- Email: [email protected]
Improvable method of fire suppressant concentration measurement: An experimental validation
Flow Measurement and Instrumentation 30 (2013) 75–80
Contents lists available at SciVerse ScienceDirect
Flow Measurement and Instrumentation journal homepage: www.elsevier.com/locate/flowmeasinst
Improvable method of fire suppressant concentration measurement: An experimental validation L.M. Zhao, X.M. Ma, J.J. Fu, D.D. Li, J. Qin n State Key Laboratory of Fire Science, University of Science and Technology of China, No.96, Jinzhai Road, Hefei, Anhui 230026, PR China
a r t i c l e i n f o
abstract
Article history: Received 8 January 2012 Received in revised form 13 September 2012 Accepted 3 January 2013 Available online 24 January 2013
An improvable method which could be used in fire suppressant concentration measurement is presented in this paper. The essential principle of the design is based on the Hagen–Poiseuille law. Gas mixture flow was driven to pass through a set of capillary tubes and pressure drop between ends of the tubes is measured. The pressure drop is a function of gas viscosity and the viscosity of binary gas mixture is actually determined by volume percentage of each component. Relationship between volume percentage and pressure drop can then be established. Results of experiments verify the theoretical analysis and show potential application prospect. & 2013 Elsevier Ltd. All rights reserved.
Keywords: Fire suppressant Concentration measurement Hagen–Poiseuille law Pressure drop Viscosity
1. Introduction The measurement of the distribution, dispersion and evaporation characteristics of a suppression agent is essential for characterizing the fire extinguishment process and certifying reliability of a fire-suppression system [1]. To ascertain the characteristic of an agent’s distribution behavior after being discharged from the fire protecting system, it is necessary to seek an appropriate way to measure the accurate and real-time concentration of fire suppressant. For another aspect, new agents are developed to meet growing demands for fire suppressant. This also calls for necessary quantitative measurement methods in order to certify whether these new agents fulfill their design objectives or not [2–4]. Moreover, real-time suppressant concentration measurement methods could also be used in the design of distribution systems and to provide technical support in the certification of new gas fire protecting systems. In the field of evaluation of fire suppression systems in aircraft engine compartments, for example, practical way to quantify the firefighting efficiency was required. Currently, the safety level for engine compartment fire protection is defined as an amount of Halon 1301 producing the minimum of a 6% volumetric concentration throughout the protected zone for a duration of one-half second. The premium is placed on the achievement of extinguishing concentrations which can be maintained for a period of time
n
Corresponding author. Tel.: þ86 0551 3601662; fax: þ86 0551 3601669. E-mail address: [email protected] (J. Qin).
0955-5986/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.flowmeasinst.2013.01.002
following extinguishment to prevent relighting on heated surfaces. This process has involved the evaluation of many chemicals and associated technologies. Better understanding of the diffusion behavior of the suppressants in the compartment, better methods of nacelle suppression could be presented and improved. Earlier work in the span of time concerned was based on actual fire testing. Results were quantified by agent weight and compared against fire extinguishment performance. These methods, however, were costly and ineffective. Generally, to examine the ability of agent to put out a fire in an unknown environment, one of the most effective ways to quantify is to measure and record concentration, distribution and duration of the agent gas in the region of interest. With the development of technology of gas concentration capture and record, agent concentration-versustime profiles illustrating distribution in the protected zone could be obtained in time. In this context, the emphasis involved finding a simpler, yet more effective, way to measure and record concentration of gaseous agents. In order to determine the characteristic of an agent’s dispersion behavior, which is the most important in determining extinguishment efficiency, it is necessary to have the means to make accurate concentration measurements on the time scales of interest. To those agents incorporated into fire-extinguishment systems which came out newly, it will also be necessary to make quantitative measurements in order to certify that the new systems are meeting their design goal. Development of the capability of making the necessary concentration measurements for the design of dispersion systems and the certification of new systems, in brief, is the focus of this effort.
76
L.M. Zhao et al. / Flow Measurement and Instrumentation 30 (2013) 75–80
Many systems used in agent gas concentration measurement were designed and constructed based on different physical or chemical fundamental principles. Researchers have developed many different kinds of devices which can be used in gas concentration measurement. For example, Yamazaki designed a gas concentration system with a venturi flowmeter and a laminar flowmeter [5]. And it showed good performance in measuring concentration of binary gas mixtures consisting of air and carbon dioxide or helium. Another measuring method was based on infrared absorption characteristics of gases. When infrared light of specific wavelength passes through mixture of several gases, the specific gas will absorb a certain amount of the infrared light. Then the amount of the outgoing light is measured and the residual processed. Concentration then can be derived from the amount of infrared loss [6]. This method needs to measure the heat loss or conduction with application of different sensors. That is to say, sensors always play an important part in the field of gas concentration measurement. Indeed, development of sensors has greatly improved the measuring accuracy. Gas pressure or pressure drop sensors which were designed based on different principles of course were developed [7]. A typical example of infrared absorption improvement was the Differential Infrared Rapid Agent Concentration Sensor (DIRRACS) build by the National Institute of Standards and Technology (NIST). It was calibrated and tested with HFC-125 [8]. Another gas concentration measurement system was built from the patent of Yanikoski [9]. The typical model was Halonyzer, which was manufactured by Pacific Scientific [10]. They improved the Statham analyzer, which was actually the predecessor of Halonyzer and put it into use in concentration measurement of Halon 1301. There are also many other measurement methods based on different or similar theories. Hot-film anemometers are most commonly used in velocity measurements. But several groups had developed probes which were capable of measuring concentration and velocity simultaneously by using two probes with very different responses [11,12]. However, the characteristic of the sensitivity and accuracy of this technique was not satisfactory. In 1992, Brown and Destefano developed the Fire Extinguishing Agent Sensor (FEAS) to provide a fast time response instrument capable of detecting the presence of an agent [13]. Similar to DIRRACS, this design was based on a technique of infrared absorption. Test results showed that response of the equipment was not perfectly flat and even a slight deviation could be significant. Another problem was its calibration of the device, which would require changing both the concentration and the rate at which the concentration was changed. As a result, it did not meet the requirements of accuracy and time response needed for a satisfactory instrument for monitoring the agent concentration. In 1993, an optical speckle technique was utilized by Oberste Lehn and Merzkirch for the indirect measurement of density fluctuations which could then be related to temperature fluctuations by assuming a constant pressure, ideal-gas flow. This technique allowed the entire flow to be analyzed simultaneously. However, it was also not appropriate for the current application since it required significant optical access and a very high degree of experiment sophistication. It would be difficult to record data in practical use. In order to find a practical way in fire agent concentration measurement, this paper discusses an improvable method. The basic principle presented in this paper is similar to that of the Statham analyzer. Tests results showed its capability of measuring concentration of CO2 in mixture of CO2 and N2 to be practical. Unlike the Statham analyzer, the concentration of fire suppressant agent was not calculated first from relative concentration
and then converted into volumetric concentration with a calibration curve. Relationship between pressure drop which was expressed with voltage signal and the volumetric concentration of CO2 in the mixture was obtained by the calibration procedure first. This result was then regarded as the standard curve. And data was stored in the computer. When carrying out a practical application, the voltage signal of pressure drop was measured, recorded and converted actually. By applying the standard curve to the data measured, the concentration could be obtained. Acquiring and processing of the data with a computer made the measurement procedure convenient and efficient. Test results were also obtained much faster and easier. This made it more timesaving to evaluate fire suppression systems.
2. Theoretical analysis 2.1. Hagen–Poiseuille equation In fluid dynamics, the Hagen–Poiseuille equation is a physical law that describes the pressure drop in a fluid flowing through a long cylindrical pipe. The assumptions of the equation are that the flow is laminar viscous and incompressible; the flow is passing through a tube with a constant circular cross-section; its length is much longer than its diameter. In standard fluid dynamics notation, it can be written as
DP ¼
8mLQ pr 4
ð1Þ
where DP is the pressure drop between the ends of cylindrical pipes; L is the length of the pipe; m is the dynamic viscosity; Q is the volumetric rate of flow of the gas; r is the radius of pipe cross area; and p is the mathematical constant. Then, with all other conditions established, DP is only the function of the dynamic viscosity of fluid. And the relationship between them is linear. Then Eq. (1) can be rewritten as
DP ¼ km
ð2Þ
where k is the instrument coefficient determined by measurement system itself. This indicates that the pressure drop between the ends of the capillary tube can be expressed with viscosity of mixture of gases and the instrument constant coefficient. 2.2. Viscosity relationship of binary gas mixture In general, viscosity of gas was obtained from experimental tests. When it was not convenient to conduct experiment or lack of database, it could be calculated with expressions as follows: P y m M i 1=2 mmix ¼ P i i 1=2 ð3Þ yi M i where mmix is the viscosity of gas mixture; yi is the mole fraction of the ith element; mi is the viscosity of the ith element; and Mi is the molecular weight of the ith element. 2.3. Calculation result The absolute viscosity of many fluids relatively does not change with the pressure but very sensitive to temperature. In fact, viscosity of gases varies widely with temperature. In gases, molecules are spare and cohesion is negligible. Thus, in gases, the exchange of momentum between layers was brought as a result of molecular movement normal to the general direction of flow, and it resists the flow. The molecular activity is known to increase with temperature, thus, the viscosity of gases will increase with temperature. The reason is a result of the consideration of the
L.M. Zhao et al. / Flow Measurement and Instrumentation 30 (2013) 75–80
77
Viscosity of mixture of CO2 and N2 (μPa·s)
17.5 Capillary tubes
17.0 Filter Heat exchanger
16.5
Gas Samples
Heating unit
16.0
Pressure
Flow meter
sensor P
Pump
Power unit Data acquisition
15.5 Data output
15.0
Data processsing Analysis unit
14.5
Fig. 2. Experimental setup and design flow of the concentration measurement system.
14.0 13.5 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Volume percentage of CO2 (100%) Fig. 1. Theoretical calculation result of viscosity of CO2 and N2 mixture.
kinetic theory. This theory indicates that gas viscosities vary directly with square root of temperature. So to eliminate effect of temperature on gas viscosity, all analysis and tests were carried out under the same temperature condition. This work designed a thermostat unit specially to keep the temperature constant (see Section 3.1). Fig. 1 is a more intuitionistic way to illustrate Eq. (3). Viscosity of CO2 and N2 was from reference [14]. It shows the theoretical calculation results of viscosity of mixed CO2 and N2 sample (293 K, 1 atm). Result shows that with the increase of CO2 volume concentration, the viscosity of mixture decreases monotonically. This makes it clear that the mixture viscosity is function of CO2 concentration. Conversely, concentration of the composition can be obtained through examining the viscosity of the mixture.
3. Experiment and results 3.1. Experiment details Experiment was designed to verify the theoretical analysis. The purpose of the experiment was to obtain the relationship between pressure drop and volumetric concentration of CO2. The experimental apparatus consisted of several parts: the power supply, the vacuum pump, the flow limiting unit, the filter unit, the heating apparatus, the data acquisition unit and the analysis unit. Gas samples were driven by the vacuum pump. The flow limiting unit enabled the samples to pass through the latter unit with controlled quantity of the volumetric rate of flow. The filter unit filtered impurities in samples and left clean and uncontaminated gas samples. Then the analysis unit transformed different concentrations into pressure difference signal. Finally, data acquisition unit acquired the signal and displayed the result. The process and experimental setup are shown in Fig. 2. The heating and analysis unit were the most important parts of the system. According to the discussions above in Section 2.3, viscosity of gas is function of temperature. Temperature is one of the most important factors that should be taken into consideration. All measurement and analysis operations must be carried out under the same temperature. Gases at different temperature were drawn into the heating unit and then were maintained at a scheduled temperature by a self-designed temperature control device (Fig. 3e) before flowing into the analysis unit. This procedure can guarantee invariance temperature, which ensures that the system can be used at different ambient temperatures. In analysis unit, capillary tubes (Fig. 3a and c) and differential pressure sensors (Fig. 3b) were introduced to generate differential
pressure. The capillary tubes generated the pressure drop and the sensors transformed it into voltage signal. Hence the pressure drop was expressed with voltage value. And in turn, by analyzing the voltage recorded with the standard curve obtained initially, the gas compositions can be obtained. The gas sample lines were seamless copper tube. The inner diameter of the sample tubes was 6 mm and the thickness was 1 mm. It was easy to bend in order to fit the measurement condition. The model of microstructure pressure sensor used in the analysis unit was serial ASDX001-D44R (Honeywell Inc.) and the model of flow meter in the limiting unit was LZB-10 (Jintai Co. Ltd). The model of pump was 2XZ-2 (Nanguang Co. Ltd.). Inner diameter of the capillary tubes in the analysis unit was 1 mm and the length was 240 mm. Thirty-six tubes were used to build a capillary bundle (Fig. 3f). Interspace between the 36 tubes and silicon rubber tubes was filled with sealant to ensure that all gas samples will flow through the capillary tubes. The selected number of tubes was aimed at enabling gases flow through the same cross-section area all the way, which guaranteed the constant volumetric rate of flow and velocity of the gas. Model of data acquisition module was RM417 (Ztic Co. Ltd.). Pressure drop voltage signals were acquired and recorded by the computer. Pictures of the main parts of the experiment setup are shown in Fig. 3. One of the most important procedures was temperature control of gas samples in analysis unit. The control device was developed and designed by researchers ourselves. It can heat gas samples with temperature from 278 K to 340 K to the same temperature 368 K. This made it possible to carry out the tests and it was not necessary to adjust the temperature of gas samples before the experiment, which was much more convenient and time-saving. Another concern was precisely controlling of flow. During the calibration tests, a meter valve was introduced between the gas sample and the flow meter. Flow flux was controlled and maintained by regulating the valve. Likewise, button on the flow meter can control and adjust the flow rate. As a matter of fact, flow meter here acts as more than an indicator, and also as a controller. With the meter valve and flow meter button all set, results showed that readings from the LZB-10 flow meter system were stable. The quantitative evidence examined to ensure stability was readings from the pressure drop sensors. Since the flow flux was the only factor that influenced the differential pressure, the results of the pressure drop can reflect the stability of the flow meter. Data processing results indicated that the uncertainty was not more than 1%. 3.2. Confirmation of operational volumetric rate of flow Tests to get relationship between flow rate and pressure drop were carried out by means of experimental facility designed above. And calibration and real tests were also performed.
78
L.M. Zhao et al. / Flow Measurement and Instrumentation 30 (2013) 75–80
240mm 10mm
Ø10mm
Silicon rubber tube Fig. 3. Picture of the main parts of the measurement system: (a) capillary tubes; (b) micro-structure pressure drop sensor; (c) capillary tubes with silicon tube package; (d) data acquisition system; and (e) heating and temperature control device.
4.0 3.5
Pressure drop signal (V)
Before the system was put into use in general application, calibration tests must be carried out first. One of the reasons was to seek the most appropriate volumetric rate of flow of the gas. This is because if the volumetric rate is too low, the pressure drop, as a result, would be quite weak. That would make it quite inconvenient to measure the exact voltage value of pressure drop and difficult to ensure the accuracy of measurement results. To meet the prerequisite of the discussion above in Section 2.1, namely the type of the flow was laminar, the volumetric rate of flow should be controlled at a certain degree properly. To establish the operational flow rate, relationship between the flow rate and pressure drop of both CO2 and N2 was measured and the results are shown in Fig. 4. According to the theoretical analysis in Section 2.1, when the volumetric rate of the gas was low, flow pattern of the gas was laminar. From Eq. (1), relationship between pressure drop and volumetric rate of the flow was linear. As shown in Fig. 4, when volumetric flow rate varied between 0.25 m3/h and 2.0 m3/h, the relationship between pressure drop and flow rate of both N2 and CO2 was also linear. If the volumetric flow rate continues to increase, the linear relation no longer exists. The test result showed a good accordance with theoretical analysis. This indicated that when the volumetric rate of gas increased to a certain extent, say 2.0–2.5 m3/h, the flow changed to turbulence flow, making it no longer applicable in linear relationship.
3.0 2.5 2.0
CO2 N2
1.5 1.0 0.5 0.0
0.5
1.0
1.5
2.0
2.5
3.0
Volumetric flow rate(m3/ h) Fig. 4. Relationship between the volumetric rate of N2 and CO2 gas and pressure drop signal.
Another obvious phenomenon was that the pressure drop of N2 at same volumetric flow rate was higher than that of CO2. This essentially was caused by the viscosity differences. Viscosity of N2 was higher than CO2 at the same temperature. As a result, the
L.M. Zhao et al. / Flow Measurement and Instrumentation 30 (2013) 75–80
difference reflected by the value of pressure drop in Fig. 4 agreed quite well with Eq. (1) in Section 2.1. According to the tests results, the volumetric rate of flow of gas samples was chosen to maintain at 1 m3/h during the subsequent experiment. With this parameter constant, Reynolds number could be controlled at 1417 for CO2 or 723 for N2. The temperature was maintained at 368 K with the heating unit. All of the necessary conditions made it available to apply the fundamental principle of the Poiseuille law. And the theory derived above could also be used to analyze the experimental results. 3.3. Test of relationship between pressure drop and volume fraction Based on the theoretical analysis, the relationship between pressure drop and volume fraction can be obtained from the fundamental theory. However, the exact viscosity value of gases is actually very difficult to obtain. For one reason, it is always affected by ambient pressure. Besides, gases of different temperature always have different viscosity (Section 2.3), making it difficult to obtain all possible data in advance. Given the situation, a calibration process was designed to determine the fundamental database. The process was as follows: first, prepare gas mixture samples of different CO2 volume concentration. In this test, 11 samples were prepared. The volume concentrations were 0%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90% and 100%. Then the gas samples were made to flow through the system and the signal output of the pressure drop caused by the 11 gas samples when they flew through the capillary tubes was measured and recorded. By analyzing the pressure drop signal of different gas samples, the relationship between pressure drop and volume fraction could be established. Before the tests, the system was powered and heating unit was preheated. Experiment was carried out when the heating unit reached the operating temperature. The vacuum pump drew gas samples into the analysis unit. The pressure sensor measured the pressure drop and transferred the signal to the computer. Then the data was acquired and processed. The relationship between pressure drop and volume fraction expressed with CO2 volume concentration was obtained (Fig. 5). 3.4. Results and discussions After the calibration procedure was completed, the standard curve should be established before the system preliminary used in analyzing the composition of gas samples. In the experiment, gas mixture consisting of CO2 and N2 was used to prove the
79
theoretical analysis in Sections 2.1 and 2.2. Pressure drop between ends of the capillary bundle was measured with microstructure pressure sensor and recorded with data acquisition software. The experiments were repeated 3 times and each pressure drop of different volume fraction was determined by its average value. All tests were carried out at the same temperature and pressure condition. The tests results were plotted in Fig. 5. It was obvious that concentration of CO2 in mixture could obviously affect the output of pressure drop signal. And the result was specific on the whole. With the increase of concentration of CO2, the pressure drop signal would decrease monotonically. Trend of the result in Fig. 5 agreed with the theoretical calculation results in Section 2.3. As mentioned above, the only difference between the gas samples was volume fraction of CO2. That indicated that in this situation, the pressure drop was only function of CO2 content in the mixture. So relationship between pressure drop and CO2 volume fraction could be established. For further analysis of the results, a fitting curve and relationship was derived. The square dot was the original measurement data and solid line curve was the fitting relationship result. Fitting process of the relationship was specific. If the volume fraction of carbon dioxide y1 in mixture was formulated as x, the fraction of nitrogen y2 was (1 x). From Eq. (3), viscosity of mixture can be rewritten as 1=2
mmix ¼
1=2
xmCO2 MCO2 þ ð1xÞmN2 M N2 1=2 xM CO2
1=2 þð1xÞM N2
ð4Þ
Further, viscosity of mixture m can be simplified down to a more compact form:
m¼
axþ b cx þ d
ð5Þ
where parameters a, b, c and d were all constants based on viscosity of CO2 and N2 at the same temperature. Substitute Eq. (5) into Eq. (2) and x can then be expressed with Dp: x¼
aDp þ b gDp þ j
ð6Þ
where a, b, g and j were also constants, according to Eq. (5). To obtain the accurate results of their value, an essential procedure of data fitting was implemented. A non-linear curve fitting method of least-square was introduced. The computational process was conducted with a fitting program and precision and accuracy of fitting results were analyzed. The fitting result of Eq. (6) was rewritten as Eq. (7). Based on the processing method, the dependence index of the fitting was 0.99805 and fitting error was 0.00027.
1.20
CO2 % ¼ Pressure drop signal (V)
1.15 Test result Fitting curve
1.10
1.05
1.00
0.95 0
10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Volume fraction of CO2 (100%) Fig. 5. Measurement results of calibration tests and fitting curve.
2:753222:34774Dp 10:55318Dp
ð7Þ
where Dp is the output signal of pressure drop. With this equation established, the relationship between pressure drop and volume fraction was definite. And the system can transform pressure drop to results of CO2 volumetric concentration. The test results, fitting results and the error between them were listed in Table 1. The error was defined as (fitting result test result)/test result. Based on the equation derived from the fitting results, the absolute value of error was less than 1%. The experimental results were found not to be exactly equal to the theoretical value. Such differences may primarily be due to departure from the pure-gas law. As has been stated, the gas samples were treated as ideal gases. This is not always tenable under real-world conditions, however. On the whole, the trends of the test results were consistent with the results of theoretical analysis.
80
L.M. Zhao et al. / Flow Measurement and Instrumentation 30 (2013) 75–80
4. Conclusion
Table 1 Test results, fitting results and fitting errors. No.
CO2 volumetric percentage (%)
Test result
Fitting result
Error (%)
1 2 3 4 5 6 7 8 9 10 11
0 10 20 30 40 50 60 70 80 90 100
1.1733 1.1589 1.1386 1.1221 1.1096 1.0850 1.0664 1.0483 1.0296 1.0055 0.9720
1.1727 1.1573 1.1413 1.1244 1.1066 1.0879 1.0681 1.0472 1.0252 1.0018 0.9769
0.05 0.14 0.24 0.20 0.27 0.27 0.16 0.10 0.43 0.37 0.50
1.0
Measurement results (100%)
0.9
An improbable method which could be used in suppressant concentration online measurement was introduced in this paper. Fundamental principle of the method was the Poiseuille law. The basic theoretical analysis was described and deduced. Experiments which were used to test and verify the analysis was conducted. Results obtained from the experiment verified the preconceived analysis. The reliability and measurement errors were satisfactory, which validated the practical feasibility of the scheme. Gas mixture of CO2 and N2 the paper used was just an example of fire suppressant. Theoretically, any mixtures of two gases with different viscosities could be analyzed in the same way. One of our next tasks is making much more precise calibration of other different kinds of gas samples which were used in fire-fighting field. In allusion to the other characteristic parameters of the system, such as response time, measurement accuracy, and measurement convenience, the improvement and majorization will also be taken into consideration.
0.8 0.7 0.6
Acknowledgments
0.5
This work is supported by the Natural Science Foundation of China (NSFC) under Grant no. 51276176. L.M. Zhao was supported by the National Basic Research Program of China (973 Program, No. 2012CB719702).
0.4 0.3 0.2 0.1 0.0 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Volume fraction of CO2 gas samples (100%) Fig. 6. Comparisons between actual concentration and measured results.
Table 2 Measurement results and full scale error. No.
Actual concentration (%)
Measurement result (%)
Full scale error (%)
1 2 3 4 5 6 7 8 9 10 11
0 10 20 30 40 50 60 70 80 90 100
0.43 8.99 21.60 31.28 38.32 51.48 60.84 69.51 78.01 88.46 101.92
0.43 1.01 1.60 1.28 1.68 1.48 0.84 0.49 1.99 1.54 1.92
To examine the reliability and accuracy of the system, another test was carried out. Eleven gas samples of mixture of CO2 and N2 with different volume concentrations vary from 0% to 100% were tested again. Comparisons between actual concentration and measured results are shown in Fig. 6. The full scale errors between them are listed in Table 2. The measured values were found to agree within 2% of full scale error with the actual concentration. It can be seen that based on the Poiseuille law and viscosity of binary gas mixture, the final calibrating equation can be established from calibration tests. Measurement results also showed the calibration was good in accordance with actual concentration. This indicated the possibility of potential applications of the method presented in this paper.
References [1] Hansberry HL. Design recommendations for fire protection of aircraft powerplant installations. Technical development report no. 31. Indianapolis, IN: Civil Aeronautics Administration Technical Development and Evaluation Center; 1943. [2] New JD, Middlesworth CM. Aircraft fire extinguishment, part III—an instrument for evaluating extinguishing systems. Technical development report no. 206. Indianapolis, IN: Civil Aeronautics Administration Technical Development and Evaluation Center; 1953. [3] Johnson AM. Fire extinguishing agent evaluation in the aircraft engine Nacelle fire test simulator. Report no. AFWAL-TR-88-2022. OH: Air Force Wright Aeronautical Labs, Wright Patterson Air Force Base; 1988. [4] Hansberry, HL. Aircraft fire extinguishment, part V—preliminary report on high-rate-discharge fire extinguishing systems for aircraft power plants. Technical development report no. 260. Indianapolis, IN: Civil Aeronautics Administration Technical Development and Evaluation Center; 1956. [5] Yamazaki Shunpei, Funaki Tatsuya, Kawashima Kenji, Kagawa Toshiharu. A concentration measurement system for binary gas mixtures using two flowmeters. Measurement Science and Technology 2007;18:2762–2768. [6] Yrjo Koskinen. Single-channel gas concentration measurement method and apparatus using a short-resonator fabry-perot interferometer. United States Patent no. 5,646,729; July 8 1997. [7] Giani Alain, Al Bayaz Asmail, Boulouz Abdel, Pascal-Delannoy Fre´de´rique, Foucaran Alain, Boyer Andre´. Thermoelectric microsensor for pressure and gas concentration measurement. Materials Science and Engineering B 2002;95:268–274. [8] Johnsson L, Mulholland W, Fraser T, Leonov I, Golubiatnikov Y. Development of a fast-response fire suppressant concentration meter. NIST Technical note 1464. Gaithersburg, MD: National Institute of Standards and Technology; 2004. [9] Yanikoski FF. Gas analysis apparatus. United States Patent Office. Patent no. 2,586,899; February 26 1952. [10] AC20-100. Advisory circular: general guidelines for measuring fireextinguishing agent concentrations in powerplant compartments; 1977. [11] Harion JL, Facre Marinet M, Camano. B. An improved method for measuring velocity and concentration by thermo-anemometry in turbulent helium–air mixture. Experiments in Fluids 1996;22:174–182. [12] Comte-Bellot G. Hot-wire anemometry. Annual Review of Fluid Mechanics 1976;8:209–231. [13] Pitts M, Mulholland W, Breuel D, Johnsson L, Chung Steve, Harris Richard H.. Real-time suppressant concentration measurement. vol. 2. Section 11. NIST SP 890; 1995. p. 401–10. [14] Wen Z, Xiaohong W, Jiguo T, Chuanguang Z. Chemical engineering principle. Dong Ying: China University of Petroleum Press; 2008 p. 266.
Contents lists available at SciVerse ScienceDirect
Flow Measurement and Instrumentation journal homepage: www.elsevier.com/locate/flowmeasinst
Improvable method of fire suppressant concentration measurement: An experimental validation L.M. Zhao, X.M. Ma, J.J. Fu, D.D. Li, J. Qin n State Key Laboratory of Fire Science, University of Science and Technology of China, No.96, Jinzhai Road, Hefei, Anhui 230026, PR China
a r t i c l e i n f o
abstract
Article history: Received 8 January 2012 Received in revised form 13 September 2012 Accepted 3 January 2013 Available online 24 January 2013
An improvable method which could be used in fire suppressant concentration measurement is presented in this paper. The essential principle of the design is based on the Hagen–Poiseuille law. Gas mixture flow was driven to pass through a set of capillary tubes and pressure drop between ends of the tubes is measured. The pressure drop is a function of gas viscosity and the viscosity of binary gas mixture is actually determined by volume percentage of each component. Relationship between volume percentage and pressure drop can then be established. Results of experiments verify the theoretical analysis and show potential application prospect. & 2013 Elsevier Ltd. All rights reserved.
Keywords: Fire suppressant Concentration measurement Hagen–Poiseuille law Pressure drop Viscosity
1. Introduction The measurement of the distribution, dispersion and evaporation characteristics of a suppression agent is essential for characterizing the fire extinguishment process and certifying reliability of a fire-suppression system [1]. To ascertain the characteristic of an agent’s distribution behavior after being discharged from the fire protecting system, it is necessary to seek an appropriate way to measure the accurate and real-time concentration of fire suppressant. For another aspect, new agents are developed to meet growing demands for fire suppressant. This also calls for necessary quantitative measurement methods in order to certify whether these new agents fulfill their design objectives or not [2–4]. Moreover, real-time suppressant concentration measurement methods could also be used in the design of distribution systems and to provide technical support in the certification of new gas fire protecting systems. In the field of evaluation of fire suppression systems in aircraft engine compartments, for example, practical way to quantify the firefighting efficiency was required. Currently, the safety level for engine compartment fire protection is defined as an amount of Halon 1301 producing the minimum of a 6% volumetric concentration throughout the protected zone for a duration of one-half second. The premium is placed on the achievement of extinguishing concentrations which can be maintained for a period of time
n
Corresponding author. Tel.: þ86 0551 3601662; fax: þ86 0551 3601669. E-mail address: [email protected] (J. Qin).
0955-5986/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.flowmeasinst.2013.01.002
following extinguishment to prevent relighting on heated surfaces. This process has involved the evaluation of many chemicals and associated technologies. Better understanding of the diffusion behavior of the suppressants in the compartment, better methods of nacelle suppression could be presented and improved. Earlier work in the span of time concerned was based on actual fire testing. Results were quantified by agent weight and compared against fire extinguishment performance. These methods, however, were costly and ineffective. Generally, to examine the ability of agent to put out a fire in an unknown environment, one of the most effective ways to quantify is to measure and record concentration, distribution and duration of the agent gas in the region of interest. With the development of technology of gas concentration capture and record, agent concentration-versustime profiles illustrating distribution in the protected zone could be obtained in time. In this context, the emphasis involved finding a simpler, yet more effective, way to measure and record concentration of gaseous agents. In order to determine the characteristic of an agent’s dispersion behavior, which is the most important in determining extinguishment efficiency, it is necessary to have the means to make accurate concentration measurements on the time scales of interest. To those agents incorporated into fire-extinguishment systems which came out newly, it will also be necessary to make quantitative measurements in order to certify that the new systems are meeting their design goal. Development of the capability of making the necessary concentration measurements for the design of dispersion systems and the certification of new systems, in brief, is the focus of this effort.
76
L.M. Zhao et al. / Flow Measurement and Instrumentation 30 (2013) 75–80
Many systems used in agent gas concentration measurement were designed and constructed based on different physical or chemical fundamental principles. Researchers have developed many different kinds of devices which can be used in gas concentration measurement. For example, Yamazaki designed a gas concentration system with a venturi flowmeter and a laminar flowmeter [5]. And it showed good performance in measuring concentration of binary gas mixtures consisting of air and carbon dioxide or helium. Another measuring method was based on infrared absorption characteristics of gases. When infrared light of specific wavelength passes through mixture of several gases, the specific gas will absorb a certain amount of the infrared light. Then the amount of the outgoing light is measured and the residual processed. Concentration then can be derived from the amount of infrared loss [6]. This method needs to measure the heat loss or conduction with application of different sensors. That is to say, sensors always play an important part in the field of gas concentration measurement. Indeed, development of sensors has greatly improved the measuring accuracy. Gas pressure or pressure drop sensors which were designed based on different principles of course were developed [7]. A typical example of infrared absorption improvement was the Differential Infrared Rapid Agent Concentration Sensor (DIRRACS) build by the National Institute of Standards and Technology (NIST). It was calibrated and tested with HFC-125 [8]. Another gas concentration measurement system was built from the patent of Yanikoski [9]. The typical model was Halonyzer, which was manufactured by Pacific Scientific [10]. They improved the Statham analyzer, which was actually the predecessor of Halonyzer and put it into use in concentration measurement of Halon 1301. There are also many other measurement methods based on different or similar theories. Hot-film anemometers are most commonly used in velocity measurements. But several groups had developed probes which were capable of measuring concentration and velocity simultaneously by using two probes with very different responses [11,12]. However, the characteristic of the sensitivity and accuracy of this technique was not satisfactory. In 1992, Brown and Destefano developed the Fire Extinguishing Agent Sensor (FEAS) to provide a fast time response instrument capable of detecting the presence of an agent [13]. Similar to DIRRACS, this design was based on a technique of infrared absorption. Test results showed that response of the equipment was not perfectly flat and even a slight deviation could be significant. Another problem was its calibration of the device, which would require changing both the concentration and the rate at which the concentration was changed. As a result, it did not meet the requirements of accuracy and time response needed for a satisfactory instrument for monitoring the agent concentration. In 1993, an optical speckle technique was utilized by Oberste Lehn and Merzkirch for the indirect measurement of density fluctuations which could then be related to temperature fluctuations by assuming a constant pressure, ideal-gas flow. This technique allowed the entire flow to be analyzed simultaneously. However, it was also not appropriate for the current application since it required significant optical access and a very high degree of experiment sophistication. It would be difficult to record data in practical use. In order to find a practical way in fire agent concentration measurement, this paper discusses an improvable method. The basic principle presented in this paper is similar to that of the Statham analyzer. Tests results showed its capability of measuring concentration of CO2 in mixture of CO2 and N2 to be practical. Unlike the Statham analyzer, the concentration of fire suppressant agent was not calculated first from relative concentration
and then converted into volumetric concentration with a calibration curve. Relationship between pressure drop which was expressed with voltage signal and the volumetric concentration of CO2 in the mixture was obtained by the calibration procedure first. This result was then regarded as the standard curve. And data was stored in the computer. When carrying out a practical application, the voltage signal of pressure drop was measured, recorded and converted actually. By applying the standard curve to the data measured, the concentration could be obtained. Acquiring and processing of the data with a computer made the measurement procedure convenient and efficient. Test results were also obtained much faster and easier. This made it more timesaving to evaluate fire suppression systems.
2. Theoretical analysis 2.1. Hagen–Poiseuille equation In fluid dynamics, the Hagen–Poiseuille equation is a physical law that describes the pressure drop in a fluid flowing through a long cylindrical pipe. The assumptions of the equation are that the flow is laminar viscous and incompressible; the flow is passing through a tube with a constant circular cross-section; its length is much longer than its diameter. In standard fluid dynamics notation, it can be written as
DP ¼
8mLQ pr 4
ð1Þ
where DP is the pressure drop between the ends of cylindrical pipes; L is the length of the pipe; m is the dynamic viscosity; Q is the volumetric rate of flow of the gas; r is the radius of pipe cross area; and p is the mathematical constant. Then, with all other conditions established, DP is only the function of the dynamic viscosity of fluid. And the relationship between them is linear. Then Eq. (1) can be rewritten as
DP ¼ km
ð2Þ
where k is the instrument coefficient determined by measurement system itself. This indicates that the pressure drop between the ends of the capillary tube can be expressed with viscosity of mixture of gases and the instrument constant coefficient. 2.2. Viscosity relationship of binary gas mixture In general, viscosity of gas was obtained from experimental tests. When it was not convenient to conduct experiment or lack of database, it could be calculated with expressions as follows: P y m M i 1=2 mmix ¼ P i i 1=2 ð3Þ yi M i where mmix is the viscosity of gas mixture; yi is the mole fraction of the ith element; mi is the viscosity of the ith element; and Mi is the molecular weight of the ith element. 2.3. Calculation result The absolute viscosity of many fluids relatively does not change with the pressure but very sensitive to temperature. In fact, viscosity of gases varies widely with temperature. In gases, molecules are spare and cohesion is negligible. Thus, in gases, the exchange of momentum between layers was brought as a result of molecular movement normal to the general direction of flow, and it resists the flow. The molecular activity is known to increase with temperature, thus, the viscosity of gases will increase with temperature. The reason is a result of the consideration of the
L.M. Zhao et al. / Flow Measurement and Instrumentation 30 (2013) 75–80
77
Viscosity of mixture of CO2 and N2 (μPa·s)
17.5 Capillary tubes
17.0 Filter Heat exchanger
16.5
Gas Samples
Heating unit
16.0
Pressure
Flow meter
sensor P
Pump
Power unit Data acquisition
15.5 Data output
15.0
Data processsing Analysis unit
14.5
Fig. 2. Experimental setup and design flow of the concentration measurement system.
14.0 13.5 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Volume percentage of CO2 (100%) Fig. 1. Theoretical calculation result of viscosity of CO2 and N2 mixture.
kinetic theory. This theory indicates that gas viscosities vary directly with square root of temperature. So to eliminate effect of temperature on gas viscosity, all analysis and tests were carried out under the same temperature condition. This work designed a thermostat unit specially to keep the temperature constant (see Section 3.1). Fig. 1 is a more intuitionistic way to illustrate Eq. (3). Viscosity of CO2 and N2 was from reference [14]. It shows the theoretical calculation results of viscosity of mixed CO2 and N2 sample (293 K, 1 atm). Result shows that with the increase of CO2 volume concentration, the viscosity of mixture decreases monotonically. This makes it clear that the mixture viscosity is function of CO2 concentration. Conversely, concentration of the composition can be obtained through examining the viscosity of the mixture.
3. Experiment and results 3.1. Experiment details Experiment was designed to verify the theoretical analysis. The purpose of the experiment was to obtain the relationship between pressure drop and volumetric concentration of CO2. The experimental apparatus consisted of several parts: the power supply, the vacuum pump, the flow limiting unit, the filter unit, the heating apparatus, the data acquisition unit and the analysis unit. Gas samples were driven by the vacuum pump. The flow limiting unit enabled the samples to pass through the latter unit with controlled quantity of the volumetric rate of flow. The filter unit filtered impurities in samples and left clean and uncontaminated gas samples. Then the analysis unit transformed different concentrations into pressure difference signal. Finally, data acquisition unit acquired the signal and displayed the result. The process and experimental setup are shown in Fig. 2. The heating and analysis unit were the most important parts of the system. According to the discussions above in Section 2.3, viscosity of gas is function of temperature. Temperature is one of the most important factors that should be taken into consideration. All measurement and analysis operations must be carried out under the same temperature. Gases at different temperature were drawn into the heating unit and then were maintained at a scheduled temperature by a self-designed temperature control device (Fig. 3e) before flowing into the analysis unit. This procedure can guarantee invariance temperature, which ensures that the system can be used at different ambient temperatures. In analysis unit, capillary tubes (Fig. 3a and c) and differential pressure sensors (Fig. 3b) were introduced to generate differential
pressure. The capillary tubes generated the pressure drop and the sensors transformed it into voltage signal. Hence the pressure drop was expressed with voltage value. And in turn, by analyzing the voltage recorded with the standard curve obtained initially, the gas compositions can be obtained. The gas sample lines were seamless copper tube. The inner diameter of the sample tubes was 6 mm and the thickness was 1 mm. It was easy to bend in order to fit the measurement condition. The model of microstructure pressure sensor used in the analysis unit was serial ASDX001-D44R (Honeywell Inc.) and the model of flow meter in the limiting unit was LZB-10 (Jintai Co. Ltd). The model of pump was 2XZ-2 (Nanguang Co. Ltd.). Inner diameter of the capillary tubes in the analysis unit was 1 mm and the length was 240 mm. Thirty-six tubes were used to build a capillary bundle (Fig. 3f). Interspace between the 36 tubes and silicon rubber tubes was filled with sealant to ensure that all gas samples will flow through the capillary tubes. The selected number of tubes was aimed at enabling gases flow through the same cross-section area all the way, which guaranteed the constant volumetric rate of flow and velocity of the gas. Model of data acquisition module was RM417 (Ztic Co. Ltd.). Pressure drop voltage signals were acquired and recorded by the computer. Pictures of the main parts of the experiment setup are shown in Fig. 3. One of the most important procedures was temperature control of gas samples in analysis unit. The control device was developed and designed by researchers ourselves. It can heat gas samples with temperature from 278 K to 340 K to the same temperature 368 K. This made it possible to carry out the tests and it was not necessary to adjust the temperature of gas samples before the experiment, which was much more convenient and time-saving. Another concern was precisely controlling of flow. During the calibration tests, a meter valve was introduced between the gas sample and the flow meter. Flow flux was controlled and maintained by regulating the valve. Likewise, button on the flow meter can control and adjust the flow rate. As a matter of fact, flow meter here acts as more than an indicator, and also as a controller. With the meter valve and flow meter button all set, results showed that readings from the LZB-10 flow meter system were stable. The quantitative evidence examined to ensure stability was readings from the pressure drop sensors. Since the flow flux was the only factor that influenced the differential pressure, the results of the pressure drop can reflect the stability of the flow meter. Data processing results indicated that the uncertainty was not more than 1%. 3.2. Confirmation of operational volumetric rate of flow Tests to get relationship between flow rate and pressure drop were carried out by means of experimental facility designed above. And calibration and real tests were also performed.
78
L.M. Zhao et al. / Flow Measurement and Instrumentation 30 (2013) 75–80
240mm 10mm
Ø10mm
Silicon rubber tube Fig. 3. Picture of the main parts of the measurement system: (a) capillary tubes; (b) micro-structure pressure drop sensor; (c) capillary tubes with silicon tube package; (d) data acquisition system; and (e) heating and temperature control device.
4.0 3.5
Pressure drop signal (V)
Before the system was put into use in general application, calibration tests must be carried out first. One of the reasons was to seek the most appropriate volumetric rate of flow of the gas. This is because if the volumetric rate is too low, the pressure drop, as a result, would be quite weak. That would make it quite inconvenient to measure the exact voltage value of pressure drop and difficult to ensure the accuracy of measurement results. To meet the prerequisite of the discussion above in Section 2.1, namely the type of the flow was laminar, the volumetric rate of flow should be controlled at a certain degree properly. To establish the operational flow rate, relationship between the flow rate and pressure drop of both CO2 and N2 was measured and the results are shown in Fig. 4. According to the theoretical analysis in Section 2.1, when the volumetric rate of the gas was low, flow pattern of the gas was laminar. From Eq. (1), relationship between pressure drop and volumetric rate of the flow was linear. As shown in Fig. 4, when volumetric flow rate varied between 0.25 m3/h and 2.0 m3/h, the relationship between pressure drop and flow rate of both N2 and CO2 was also linear. If the volumetric flow rate continues to increase, the linear relation no longer exists. The test result showed a good accordance with theoretical analysis. This indicated that when the volumetric rate of gas increased to a certain extent, say 2.0–2.5 m3/h, the flow changed to turbulence flow, making it no longer applicable in linear relationship.
3.0 2.5 2.0
CO2 N2
1.5 1.0 0.5 0.0
0.5
1.0
1.5
2.0
2.5
3.0
Volumetric flow rate(m3/ h) Fig. 4. Relationship between the volumetric rate of N2 and CO2 gas and pressure drop signal.
Another obvious phenomenon was that the pressure drop of N2 at same volumetric flow rate was higher than that of CO2. This essentially was caused by the viscosity differences. Viscosity of N2 was higher than CO2 at the same temperature. As a result, the
L.M. Zhao et al. / Flow Measurement and Instrumentation 30 (2013) 75–80
difference reflected by the value of pressure drop in Fig. 4 agreed quite well with Eq. (1) in Section 2.1. According to the tests results, the volumetric rate of flow of gas samples was chosen to maintain at 1 m3/h during the subsequent experiment. With this parameter constant, Reynolds number could be controlled at 1417 for CO2 or 723 for N2. The temperature was maintained at 368 K with the heating unit. All of the necessary conditions made it available to apply the fundamental principle of the Poiseuille law. And the theory derived above could also be used to analyze the experimental results. 3.3. Test of relationship between pressure drop and volume fraction Based on the theoretical analysis, the relationship between pressure drop and volume fraction can be obtained from the fundamental theory. However, the exact viscosity value of gases is actually very difficult to obtain. For one reason, it is always affected by ambient pressure. Besides, gases of different temperature always have different viscosity (Section 2.3), making it difficult to obtain all possible data in advance. Given the situation, a calibration process was designed to determine the fundamental database. The process was as follows: first, prepare gas mixture samples of different CO2 volume concentration. In this test, 11 samples were prepared. The volume concentrations were 0%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90% and 100%. Then the gas samples were made to flow through the system and the signal output of the pressure drop caused by the 11 gas samples when they flew through the capillary tubes was measured and recorded. By analyzing the pressure drop signal of different gas samples, the relationship between pressure drop and volume fraction could be established. Before the tests, the system was powered and heating unit was preheated. Experiment was carried out when the heating unit reached the operating temperature. The vacuum pump drew gas samples into the analysis unit. The pressure sensor measured the pressure drop and transferred the signal to the computer. Then the data was acquired and processed. The relationship between pressure drop and volume fraction expressed with CO2 volume concentration was obtained (Fig. 5). 3.4. Results and discussions After the calibration procedure was completed, the standard curve should be established before the system preliminary used in analyzing the composition of gas samples. In the experiment, gas mixture consisting of CO2 and N2 was used to prove the
79
theoretical analysis in Sections 2.1 and 2.2. Pressure drop between ends of the capillary bundle was measured with microstructure pressure sensor and recorded with data acquisition software. The experiments were repeated 3 times and each pressure drop of different volume fraction was determined by its average value. All tests were carried out at the same temperature and pressure condition. The tests results were plotted in Fig. 5. It was obvious that concentration of CO2 in mixture could obviously affect the output of pressure drop signal. And the result was specific on the whole. With the increase of concentration of CO2, the pressure drop signal would decrease monotonically. Trend of the result in Fig. 5 agreed with the theoretical calculation results in Section 2.3. As mentioned above, the only difference between the gas samples was volume fraction of CO2. That indicated that in this situation, the pressure drop was only function of CO2 content in the mixture. So relationship between pressure drop and CO2 volume fraction could be established. For further analysis of the results, a fitting curve and relationship was derived. The square dot was the original measurement data and solid line curve was the fitting relationship result. Fitting process of the relationship was specific. If the volume fraction of carbon dioxide y1 in mixture was formulated as x, the fraction of nitrogen y2 was (1 x). From Eq. (3), viscosity of mixture can be rewritten as 1=2
mmix ¼
1=2
xmCO2 MCO2 þ ð1xÞmN2 M N2 1=2 xM CO2
1=2 þð1xÞM N2
ð4Þ
Further, viscosity of mixture m can be simplified down to a more compact form:
m¼
axþ b cx þ d
ð5Þ
where parameters a, b, c and d were all constants based on viscosity of CO2 and N2 at the same temperature. Substitute Eq. (5) into Eq. (2) and x can then be expressed with Dp: x¼
aDp þ b gDp þ j
ð6Þ
where a, b, g and j were also constants, according to Eq. (5). To obtain the accurate results of their value, an essential procedure of data fitting was implemented. A non-linear curve fitting method of least-square was introduced. The computational process was conducted with a fitting program and precision and accuracy of fitting results were analyzed. The fitting result of Eq. (6) was rewritten as Eq. (7). Based on the processing method, the dependence index of the fitting was 0.99805 and fitting error was 0.00027.
1.20
CO2 % ¼ Pressure drop signal (V)
1.15 Test result Fitting curve
1.10
1.05
1.00
0.95 0
10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Volume fraction of CO2 (100%) Fig. 5. Measurement results of calibration tests and fitting curve.
2:753222:34774Dp 10:55318Dp
ð7Þ
where Dp is the output signal of pressure drop. With this equation established, the relationship between pressure drop and volume fraction was definite. And the system can transform pressure drop to results of CO2 volumetric concentration. The test results, fitting results and the error between them were listed in Table 1. The error was defined as (fitting result test result)/test result. Based on the equation derived from the fitting results, the absolute value of error was less than 1%. The experimental results were found not to be exactly equal to the theoretical value. Such differences may primarily be due to departure from the pure-gas law. As has been stated, the gas samples were treated as ideal gases. This is not always tenable under real-world conditions, however. On the whole, the trends of the test results were consistent with the results of theoretical analysis.
80
L.M. Zhao et al. / Flow Measurement and Instrumentation 30 (2013) 75–80
4. Conclusion
Table 1 Test results, fitting results and fitting errors. No.
CO2 volumetric percentage (%)
Test result
Fitting result
Error (%)
1 2 3 4 5 6 7 8 9 10 11
0 10 20 30 40 50 60 70 80 90 100
1.1733 1.1589 1.1386 1.1221 1.1096 1.0850 1.0664 1.0483 1.0296 1.0055 0.9720
1.1727 1.1573 1.1413 1.1244 1.1066 1.0879 1.0681 1.0472 1.0252 1.0018 0.9769
0.05 0.14 0.24 0.20 0.27 0.27 0.16 0.10 0.43 0.37 0.50
1.0
Measurement results (100%)
0.9
An improbable method which could be used in suppressant concentration online measurement was introduced in this paper. Fundamental principle of the method was the Poiseuille law. The basic theoretical analysis was described and deduced. Experiments which were used to test and verify the analysis was conducted. Results obtained from the experiment verified the preconceived analysis. The reliability and measurement errors were satisfactory, which validated the practical feasibility of the scheme. Gas mixture of CO2 and N2 the paper used was just an example of fire suppressant. Theoretically, any mixtures of two gases with different viscosities could be analyzed in the same way. One of our next tasks is making much more precise calibration of other different kinds of gas samples which were used in fire-fighting field. In allusion to the other characteristic parameters of the system, such as response time, measurement accuracy, and measurement convenience, the improvement and majorization will also be taken into consideration.
0.8 0.7 0.6
Acknowledgments
0.5
This work is supported by the Natural Science Foundation of China (NSFC) under Grant no. 51276176. L.M. Zhao was supported by the National Basic Research Program of China (973 Program, No. 2012CB719702).
0.4 0.3 0.2 0.1 0.0 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Volume fraction of CO2 gas samples (100%) Fig. 6. Comparisons between actual concentration and measured results.
Table 2 Measurement results and full scale error. No.
Actual concentration (%)
Measurement result (%)
Full scale error (%)
1 2 3 4 5 6 7 8 9 10 11
0 10 20 30 40 50 60 70 80 90 100
0.43 8.99 21.60 31.28 38.32 51.48 60.84 69.51 78.01 88.46 101.92
0.43 1.01 1.60 1.28 1.68 1.48 0.84 0.49 1.99 1.54 1.92
To examine the reliability and accuracy of the system, another test was carried out. Eleven gas samples of mixture of CO2 and N2 with different volume concentrations vary from 0% to 100% were tested again. Comparisons between actual concentration and measured results are shown in Fig. 6. The full scale errors between them are listed in Table 2. The measured values were found to agree within 2% of full scale error with the actual concentration. It can be seen that based on the Poiseuille law and viscosity of binary gas mixture, the final calibrating equation can be established from calibration tests. Measurement results also showed the calibration was good in accordance with actual concentration. This indicated the possibility of potential applications of the method presented in this paper.
References [1] Hansberry HL. Design recommendations for fire protection of aircraft powerplant installations. Technical development report no. 31. Indianapolis, IN: Civil Aeronautics Administration Technical Development and Evaluation Center; 1943. [2] New JD, Middlesworth CM. Aircraft fire extinguishment, part III—an instrument for evaluating extinguishing systems. Technical development report no. 206. Indianapolis, IN: Civil Aeronautics Administration Technical Development and Evaluation Center; 1953. [3] Johnson AM. Fire extinguishing agent evaluation in the aircraft engine Nacelle fire test simulator. Report no. AFWAL-TR-88-2022. OH: Air Force Wright Aeronautical Labs, Wright Patterson Air Force Base; 1988. [4] Hansberry, HL. Aircraft fire extinguishment, part V—preliminary report on high-rate-discharge fire extinguishing systems for aircraft power plants. Technical development report no. 260. Indianapolis, IN: Civil Aeronautics Administration Technical Development and Evaluation Center; 1956. [5] Yamazaki Shunpei, Funaki Tatsuya, Kawashima Kenji, Kagawa Toshiharu. A concentration measurement system for binary gas mixtures using two flowmeters. Measurement Science and Technology 2007;18:2762–2768. [6] Yrjo Koskinen. Single-channel gas concentration measurement method and apparatus using a short-resonator fabry-perot interferometer. United States Patent no. 5,646,729; July 8 1997. [7] Giani Alain, Al Bayaz Asmail, Boulouz Abdel, Pascal-Delannoy Fre´de´rique, Foucaran Alain, Boyer Andre´. Thermoelectric microsensor for pressure and gas concentration measurement. Materials Science and Engineering B 2002;95:268–274. [8] Johnsson L, Mulholland W, Fraser T, Leonov I, Golubiatnikov Y. Development of a fast-response fire suppressant concentration meter. NIST Technical note 1464. Gaithersburg, MD: National Institute of Standards and Technology; 2004. [9] Yanikoski FF. Gas analysis apparatus. United States Patent Office. Patent no. 2,586,899; February 26 1952. [10] AC20-100. Advisory circular: general guidelines for measuring fireextinguishing agent concentrations in powerplant compartments; 1977. [11] Harion JL, Facre Marinet M, Camano. B. An improved method for measuring velocity and concentration by thermo-anemometry in turbulent helium–air mixture. Experiments in Fluids 1996;22:174–182. [12] Comte-Bellot G. Hot-wire anemometry. Annual Review of Fluid Mechanics 1976;8:209–231. [13] Pitts M, Mulholland W, Breuel D, Johnsson L, Chung Steve, Harris Richard H.. Real-time suppressant concentration measurement. vol. 2. Section 11. NIST SP 890; 1995. p. 401–10. [14] Wen Z, Xiaohong W, Jiguo T, Chuanguang Z. Chemical engineering principle. Dong Ying: China University of Petroleum Press; 2008 p. 266.