Measuring the flammability of mixtures of combustible gases and oxygen

FIELD OF THE INVENTION

0001The present invention relates to the measurement of combustible gases and oxygen mixtures in atmospheres having the potential for a fire or explosion, and in particular to temperature measurements of diffusion flames within samples of combustible mixtures.

BACKGROUND OF THE INVENTION

0002<u>Description of Related Art</u>

0003Measurement of combustibles gases and oxygen in industrial processes can prevent fires and explosions. The National Fire Protection Association NFPA-86 Standard for Ovens and Furnaces (NFPA-86) defines the safety requirements for flammability measurements in industrial processes that use combustible substances, typically solvents, for printing, painting, coating or treating various materials, which are then dried in ovens. For combustible gases, this is achieved by keeping the concentration under the Lower Flammable Limit, or LFL, below which the combustible gas cannot propagate a flame. A margin of safety is often required, so that the limit to an industrial operation is often expressed as a percentage of the LFL, for example 50 %LFL. For oxygen, this means keeping the concentration, by a safe margin, under the limiting oxygen concentration, or LOC, below which there is insufficient oxygen for an explosion, for example 3% oxygen below the LOC.

0004Efforts to reduce emissions of volatile organic compounds (VOCs) from industrial sources have resulted in the widespread use of thermal oxidizers. VOCs that were once released to the atmosphere are now captured and transported to an oxidizer to be destroyed. Since VOCs are combustible, they can form explosive mixtures in air. The thermal oxidizer, as well as the process ductwork which transports the VOCs to the oxidizer, must be protected. The oxidizer is a near-certain source of ignition to an explosive mixture, and the oxidizer's function requires sufficient oxygen to support combustion, so the primary method of protection is to limit the concentration of VOCs to a level safely below the LFL.

0005Recent improvements in manufacturing techniques have led to increases in throughput, which can mean that higher volumes of combustibles must be handled safely. It is not uncommon for a particular process with a long history of safe operation to experience an incident after a seemingly unrelated change in speed, formulation, or temperature causes an unintentional increase in the concentration of combustibles.

0006The drive to become energy efficiency focuses on reducing the amount of fuel used for heating, which is sometimes achieved by reducing ventilation, and in so doing, results in higher concentrations of combustibles. The need to measure flammability accurately under a wide range of conditions is more important than ever.

0007Process conditions can vary widely. The types of combustible gases and their proportions, the oxygen concentration, water vapor content, and other gases in the process might be unpredictable, or may vary over a wide range. The conditions during normal operation might be well defined, but under a potentially hazardous process upset, these conditions may be significantly different than what was originally anticipated. An analyzer capable of giving a reliable alarm under the widest range of potential circumstances is needed.

0008For example, a kiln in which ceramics are fired have rarely experienced problems in the past, but now kilns may have more complex parts with greater surface areas, using shorter cycle times, and more volatile binders. The concentration of combustibles can now reach very high concentrations. Furthermore, the process can generate an extremely wide range of combustible vapors - many different hydrocarbon species of varying molecular weights, plus hydrogen or carbon monoxide, depending upon the kiln conditions at the time the vapors are generated. In some cases more than forty different hydrocarbon species can be detected within the kiln atmosphere. The kiln temperature and oxygen range is also extremely broad. It is difficult to predict the type of combustible vapor that may be present, so an analyzer with a universal response to combustible gases that also can tolerate wide ranges of temperature and oxygen is needed.

0009NFPA-86 describes the characteristics of analyzers used in the measurement of combustible gases in industrial processes. These include catalytic bead sensor, infrared absorption analyzer (IR), flame temperature analyzer (FTA), and flame ionization detector (FID). It makes clear the need to choose an analyzer with the correct characteristics. The analyzer should be fast, accurate and reliable.

0010"The choice of the appropriate type depends on factors such as the solvent composition, the need for calibration of more than one solvent, the necessary response time of the measurement system, and the ability to handle factors such as contaminants and oxygen content."

0011Control Instruments Corporation developed the FTA, or flame temperature analyzer. One embodiment of this analyzer is described in <patcit id="pcit0001" dnum="US5053200A"><text>U.S. patent 5,053,200</text></patcit>. It is well suited for flammability measurements under a wide range of conditions. It uses a temperature detector to measure the heat produced by a small "sensing flame." The sensing flame is typically fueled by a small, continuous flow of propane or hydrogen fuel. A sample is extracted from the atmosphere being monitored. The sensing flame sits in a combustion chamber that is completely filled with the sample. A temperature detector is positioned a short distance above the sensing flame, where it can sense the heat produced by the sensing flame. When combustible gases are absent from the sample, the temperature detector receives heat only from fuel combustion: it is at the "zero temperature as in 0% LFL." When combustible gases appear in the sample, they diffuse into the sensing flame, are burned, release heat, and produce a temperature rise in the temperature detector that is directly proportional to the flammability of the sample. The FTA analyzer is calibrated with a gas having a known flammability, so that resulting rise in temperature above the zero temperature, known as the "span temperature," is determined. Following calibration, the flammability of a sample can be calculated from the zero and span temperatures.

0012Another FTA, for detecting combustible gases in an atmosphere, is disclosed in <patcit id="pcit0002" dnum="GB778774A"><text>GB 778 774</text></patcit>.

0013A significant advantage of the FTA analyzer is its uniform response to a wide variety of combustible gases and vapors. This is extremely useful when the combustible gases, or their relative proportions in a mixture, are unknown. Experiments have shown that this is achieved through diffusion of combustible gases from the sample into the sensing flame.

0014One of the few disadvantages of the FTA has been the effect that varying concentrations of oxygen in the sample can have on the zero temperature. The sensing flame requires oxygen from the sample. As the sample's oxygen concentration decreases, the flame begins to starve. It grows larger and changes shape. This in turn produces a false signal, called the oxygen effect, which is equivalent to a shift in the zero temperature of about +1% LFL for every 1% decrease in the sample's oxygen concentration. Below a 10% sample oxygen concentration, the flame begins to destabilize, and the error signal declines somewhat. At about 6% the sensing flame is extinguished, which is called flameout. The FTA oxygen effect is equivalent to a change in zero temperature because it is independent of the concentration of combustible gases in the sample.

0015An important characteristic of the FTA's sensing flame is its inherently failsafe design. The "zero" temperature of approximately 500°C disposes the device to readily reveal faults. For example, a failure of flow to the analyzer, such as can occur from a clogged sample tube, will cause flameout from oxygen starvation and produce a significant negative deflection in the measurement circuit that is readily distinguished from valid measurements. A fault signal can therefore signal this condition to prevent mis-interpretation of the analyzer reading. To maintain this failsafe characteristic, the design of the sensing flame must ideally be such that oxygen starvation causes only positive deflections in measurement right up to the point of flameout, at which point a sudden and complete destabilization of the flame, and flameout, occurs. In this way, errors may cause exaggerated readings, but can never under-report the concentration of combustibles.

0016The accuracy of measurement typically required of a flammability analyzer by industry standards is +/-3%LFL. An improvement to the FTA which would reduce the oxygen effect below this accuracy limit is therefore desirable. Because they both exceed the limit to accuracy, and also underreport the concentration of combustibles, errors more negative than -3%LFL are particularly undesirable.

0017One attempt by Control Instruments Corporation to improve the FTA is called air dilution. Air dilution is a sample preparation technique, whereby equal amounts of air and sample are mixed together to form an air-sample mixture that is then drawn into the analyzer. By using equal amounts of air and sample, known as a 1:1 air dilution ratio, the air-sample mixture will always contain at least 10% oxygen, which prevents instability and flameout. However, air dilution by itself does not improve the oxygen effect. Even though the magnitude of the oxygen effect is halved, so too is the concentration of combustible gases. So the error from the oxygen effect remains in equal proportion to the flammability signal. The analyzer's indicated reading will still have about the same error, expressed in terms of %LFL. Furthermore, the signal to noise ratio is unfavorably decreased, and the sample preparation apparatus introduces a new source of error. Unintentional variations in the ratio of air to sample can cause significant errors in the measurement.

0018Due to the limitations given above, air dilution and support air techniques have been used only where absolutely necessary to prevent flameout, but so far have not led to a significant improvement in performance of the analyzer.

SUMMARY OF THE INVENTION

0019The invention is directed to a gas analyzing method according to claim 1.

0020Briefly described, the invention is a novel Flame Temperature Analyze (FTA) method capable of measuring both the concentration of combustible gases and the concentration of oxygen in a gas sample. In addition, the FTA of this invention significantly reduces the "oxygen effect" error, exhibited by previous FTA's when the oxygen content of the sample is low. The FTA of this invention functions in part by supplying the sensing, diffusion flame with a mixture that includes a fuel and an oxidant.

0021The addition of an appropriate amount of oxidant to the sensing flame mixture allows the flame to burn with an inner ignition or combustion zone that stabilizes the main reaction zone. As in a conventional FTA, the main reaction zone is where combustion of the flame fuel with the oxidant from the sample occurs as the fuel diffuses into the sample. In the inner combustion zone of the FTA, of the present invention, the flame fuel is partially burned with the oxidant supplied to it, before reaching and providing stability to the outer, main reaction zone. As in conventional FTAs, the heat produced by the main reaction zone depends on the amount of combustible gases, including oxidant, in the sample. A temperature detector measures the heat produced by this part of the sensing flame, enabling the concentration of combustible gases to be calculated.

0022In the FTA of this invention, the oxygen content of the fuel mixture is maintained at around 10% by volume. This ratio of oxygen ensures that the inner ignition zone only burns part of the fuel while pxe-heating the remainder of the fuel, thereby providing the main combustion zone with improved stability in low sample oxygen situations.

0023A sample is drawn from the atmosphere being monitored, typically by suction from a pump or aspirator. The sample fills the flame cell. The sensing flame is fed by a source of fuel and oxidant through a burner tube, An igniter, typically a spark discharge to the grounded burner, lights the sensing flame. A temperature detector, typically a thermocouple, is positioned directly above the sensing flame. The temperature detector measures the heat produced by the sensing flame. A measuring circuit amplifies and conditions the temperature detector signal into a flammability reading that represents the concentration of oxygen and/or combustibles gases, or both.

0024Preferably, the sensing diffusion flame is further stabilized by including a neutral gas in the flame fuel mixture. The flame fuel mixture may consist of air and hydrogen in nearly equal proportions. This mixture has approximately 10% oxygen, 40% Nitrogen and 50% hydrogen, suitable for creating the inner, partial combustion zone that provides stability to the sensing flame.

0025Preferably, the flame temperature analyzer is further stabilized by increasing the initial "zero" flame temperature from a usual value of 500°C to 625°C in conjunction with air dilution of the sample. The sample : air dilution ratio in the preferred embodiment of this version of the invention is a one-to-one.

0026The oxygen content of the sample can be determined by monitoring the variation in the sensing flame's temperature as the oxygen content of the flame fuel mixture is varied.

BRIEF DESCRIPTION OF THE DRAWING

0027<figref idref="f0001">FIG. 1</figref> is a side, elevational, partial cross-sectional view of an analyzer which can be used in the method of the invention.

0028<figref idref="f0001">FIG. 2</figref> is a front, elevational, partial cross-sectional view of an analyzer which can be used in the method of the invention.

0029<figref idref="f0002">FIGS. 3a-3c</figref> are exploded views of the flame cell of an analyzer which can be used in the method of the invention.

0030<figref idref="f0002 f0003">FIGS. 4a-4c</figref> are schematic flow diagrams. <figref idref="f0002">FIG. 4a</figref> represents prior art FTA analyzer such as that disclosed under <patcit id="pcit0003" dnum="US5053200A"><text>U.S. patent 5,053,200</text></patcit>. <figref idref="f0003">FIG. 4b</figref> represents a prior art FID such as in <patcit id="pcit0004" dnum="US3767363A"><text>U.S. patent 3,767,363</text></patcit>. <figref idref="f0003">FIG. 4c</figref> represents a flame which can be used in the method of the invention.

0031<figref idref="f0004">FIG. 5</figref> is a system block diagram.

0032<figref idref="f0005 f0006 f0007">FIG. 6a-6f</figref> are graphs, and <figref idref="f0007">FIG 6g</figref> is a table, of the FTA analyzer reading versus sample oxygen concentration for various fuels.

0033<figref idref="f0008">FIGS. 7a-7c</figref> are schematic representations of flames used in FTA and FID analyzers. <figref idref="f0008">FIG. 7a</figref> represents the flame in a prior art FTA analyzer such as that disclosed under <patcit id="pcit0005" dnum="US5053200A"><text>U.S. patent 5,053,200</text></patcit>. <figref idref="f0008">FIG. 7b</figref> represents a prior art FID such as in <patcit id="pcit0006" dnum="US3767363A"><text>U.S. patent 3,767,363</text></patcit>. <figref idref="f0008">FIG. 7c</figref> represents a flame which can be used in the method of the invention.

DETAILED DESCRIPTION

0034During the course of this description, like numbers will be used to identify like elements according to the different views of an apparatus which can be used in the method of the invention.

0035An apparatus which can be used in the method of the invention is illustrated in <figref idref="f0001">FIG. 1</figref>. The analyzer 10 mounts to the process duct 12 in order to measure the gases therein. The analyzer 10 contains a sample flow system 28, which draws a portion of the gases flowing through process duct 12 through sample tubing 26 into flame cell assembly 16. The gases fill an internal chamber formed by flame cell block 20 and burner plate 18 that contains a sensing flame 48 and temperature detector 52. The heat produced by combustion of these gases in the sensing flame 48 causes an electrical signal in temperature detector 52. The temperature detector 52 is connected to controller 14 by lead wires that carry an electrical signal. This electrical signal is amplified by a measuring circuit 54, converted to a digital signal by an analog to digital converter 76, and conditioned by microprocessor 78 for display as a %LFL reading or %Oxygen reading on a signal meter 84. Microprocessor 78 can compare these readings to pre-programmed alarm settings and generate warning or danger alarms that activate alarm relays 82 which can be used to notify personnel of potentially dangerous conditions, and can be used to automatically and safely stop the process being monitored.

0036A sample flow system as described in detail in <patcit id="pcit0007" dnum="US5053200A"><text>U.S. patent 5,053,200</text></patcit> can be used to reliably extract a portion of the gases from the process duct 12 and fill the inner chamber of the flame cell assembly 16. Air from compressed air supply 93 passes into the aspirator 94 located within the sample flow system 28 and is ejected through the exhaust tubing 96. Expansion of the compressed air creates suction at the aspirator 94 that causes gases to be drawn from the process duct 12 into the sample tubing 26 through the sample filter 62 and sample flame arrestor 64 into the flame cell, past the sensing flame, and then through the exhaust flame arrester 66 and exhaust filter 68 and out of the flame cell assembly. The gases then pass through a sample flow orifice 70 and to the aspirator 94 where they are ejected from the aspirator out though the exhaust tubing 96 and back to the process duct 12.

0037A pressure transducer 72 monitors the pressure drop across sample flow orifice 70. The electrical signal produced by the pressure transducer 72 is amplified by the transducer amplifier 74 located within controller 14, is converted into a digital signal by analog to digital converter 76, and compared by microprocessor 78 to pre-programmed settings that represent a minimum acceptable sample flow rate. If the flow rate falls below the minimum acceptable setting, the microprocessor 78 activates a fault using alarm relays 82 to notify personnel of the fault condition.

0038The flame cell assembly consists of a burner plate assembly 17 attached to a flame cell block 20 by flame cell mounting bolts 22. A burner plate o-ring 44 seals the joint between the burner plate assembly 17 and flame cell block 20 to prevent leaks that might allow sample gases to escape or ambient gases to penetrate the flame cell. The flame cell assembly is likewise mounted to the sample flow system by flame cell mounting bolts 22. The entry for the sample gases and exit for exhaust gases from the flame cell are sealed by the sample inlet o-ring 24 and exhaust o-ring 86 respectively. The sample filter 62 fits inside the sample flame arrestor 64 which screws into the flame cell block 20 and is sealed by a sample flame arrestor cap 65. Likewise the exhaust filter 68 fits inside the exhaust flame arrestor 66 and is sealed by the exhaust flame arrestor cap 67. An observation window 49 is inserted in a threaded hole in the flame cell block 20 to provide the means to observe the sensing flame, or spark ignition, without having to remove the flame cell mounting bolts 22.

0039The burner plate assembly 17 holds the temperature detector 52 and burner 46. A spark plug 98 passes through the burner plate 18 in the vicinity of the burner 46, and is positioned so that a high voltage can be applied to the spark plug 98 so that it discharges to the burner 46 and ignites the sensing flame 48. Fuel from the fuel supply 30 passes through the fuel pressure control valve 32 and then through the fuel orifice 34 before passing through bored holes in the burner plate 18 and on the to the burner 46. The pressure of the fuel pressure control valve 32 in conjunction with the size of the fuel orifice 34 determines the rate of flow of fuel to the burner 46. Likewise, oxidant, preferably air, from an oxidant supply 36 passes through the oxidant pressure control valve 38 and then through the oxidant orifice 40 before passing through bored holes in the burner plate 18 and on to the burner 46. The pressure of the oxidant pressure control valve 38 in conjunction with the size of the oxidant orifice 40 determines the rate of flow of oxidant to the burner 46.

0040The measuring circuit conditions the voltage signal from the temperature detector. The temperature of the detector with no combustibles in the sample is known as the "zero." A typical zero temperature is 500 degrees Celsius, measured using a sheathed thermocouple about 1 cm above the flame. The difference between the zero and the temperature of the detector when the sample contains a known concentration of test gas is called the "span." A typical span temperature is approximately 100 degrees Celsius for a 50% LFL mixture, or about 2 degrees Celsius per percent LFL. When the flame cell has a sample of unknown flammability, the measuring circuit can interpolate the reading from the zero and span, and display the flammability of the sample in terms of %LFL. For example, when the zero is 500°C and the span is 2°C per %LFL, a temperature of 520°C would be interpreted as a concentration of 10 %LFL in the sample.

0041By flame temperature, applicants mean the flame temperature as measured. This temperature is not absolute, but depends upon the way in which the thermocouple is positioned relative to the flame. It can for example be the flame temperature as measured by a thermocouple located a short distance above the flame. Alternatively, the flame temperature may be measured by a thermocouple attached to the burner. The important aspect of these experimental temperature measurements is self consistency.

0042In the preferred embodiment, with the flame temperature measured with a thermocouple above the flame, the zero temperature may be 625 Deg C, whereas with an alternate placement of the thermocouple, for example to the tip of the burner, or perhaps twice the normal distance from the flame, the flame temperature might be lower, or higher, in absolute terms. The optimal performance would still be obtained with the same fuel flow rates and oxygen concentrations, but the absolute temperature measured might be different. Our measurements should not be confused with what combustion engineers refer to as "adiabatic flame temperature" which is a theoretical concept based upon calculation from thermodynamic data.

0043When the sensing flame is fed pure fuel, an oxygen effect of approximately - 2°C per percent by volume oxygen is observed. For example, if the zero reading is 500°C with air (20.9 % by volume oxygen), when the oxygen concentration of the sample decreased by 10% by volume oxygen, the temperature increases to 520°C. The oxygen effect of +20°C is equivalent to +10%LFL. Oxygen effects as high as +15%LFL can be observed. With the addition of oxidant to the fuel, the oxygen effect is greatly reduced, and the oxygen effect, over the entire range of oxygen concentrations that will still support the sensing flame, can be held to less than 3 %LFL.

0044<figref idref="f0005 f0006 f0007">Figures 6a through 6f</figref> are graphs of the oxygen effect in a FTA analyzer for various fuel types. In these graphs, the independent variable is the oxygen concentration in the sample, which was varied from 20.9 percent by volume, the normal atmospheric concentration, down to the point of flameout, where the sensing flame could no longer be sustained. The dependent variable is the analyzer's flammability reading.

0045<figref idref="f0005">Figure 6a</figref> shows the oxygen effect on the zero reading, and also on the span reading, for the prior art FTA analyzer.

0046<figref idref="f0005">Figure 6b</figref> shows the oxygen effect on the zero reading for a FTA analyzer for an initial zero flame temperature of 500°C, and also for an initial zero flame temperature of 625°C. Both flames remain lit down to about 5% by volume oxygen. The 500°C flame has a maximum zero effect of about +12%LFL and exhibits no negative readings. The 625°C flame has a maximum zero effect of about -27%LFL, with most of the zero effect occurring below 10% by volume oxygen.

0047<figref idref="f0006">Figure 6c</figref> shows the oxygen effect for several hydrogen-oxygen fuel mixtures. The oxygen effect for pure (100%) hydrogen fuel is also shown for reference. Hydrogen-oxygen fuel mixtures exhibit less than half the oxygen effect compared to pure hydrogen.

0048<figref idref="f0006">Figure 6d</figref> shows the oxygen effect for several hydrogen-nitrogen mixtures. The oxygen effect for pure (100%) hydrogen fuel is also shown for reference. The hydrogen-nitrogen fuels exhibit a significant zero effect which includes undesirable negative readings.

0049<figref idref="f0007">Figure 6e</figref> shows the oxygen effect for several hydrogen-air fuel mixtures. Hydrogen-air fuels of 46% Air to 50% Air have greatly reduced oxygen effect and only slight negative readings.

0050<figref idref="f0007">Figure 6f</figref> shows the use of improved fuel types to improve the performance of air dilution. The oxygen effect, when air dilution in a 1:1 ratio of air to sample is used to allow operation down to 0% by volume oxygen, is, for the 625°C hydrogen flame, less than half that of the 500°C hydrogen flame. The use of hydrogen-air fuel, with an air dilution in a 2:1 ratio of sample to air, allows operation down to 0% oxygen with a maximum error less than 3%LFL. The improved analyzer can operate down to 0% by volume oxygen without flameout and without negative readings.

0051<figref idref="f0007">Figure 6g</figref> is a table that shows both the oxygen effect at 10% oxygen, and the change in signal strength, for a series of different initial flame temperatures.

0052If measurements are made with oxidant and without, the two measurements can be compared to one another, and the oxygen concentration can be derived. This allows the sensing flame to make oxygen measurements independent of the concentration of combustibles, and combustible measurements independent of the concentration of oxygen.

0053The pressure of the oxidant pressure control valve 38 may be varied by means of control elements 80 attached to microprocessor 78, so that the oxidant flow rate can be varied in order to determine the concentration of oxygen in the sample gases.

0054The invention therefore provides a method for measuring the concentration of combustible gas and the concentration of oxygen in a mixture, whereby a temperature detector measures the heat produced by a sensing flame that is formed by a mixture of fuel and an oxidant emitted from a burner.

0055The oxidant can be air from a compressed air supply, or oxygen from a compressed oxygen cylinder. The oxidant may also be a mixture of air or oxygen and an inert gas such as nitrogen, argon, helium or the like. The concentration of oxygen must be sufficient to create a stable combustion zone in the vicinity of the burner tip. According to the invention, this oxygen concentration is optimally 10% by volume. The range from approximately 9.4% to 10.5% by volume oxygen gives good results. The system which delivers oxygen to the burner may therefore be designed to deliver 10% by volume oxygen, and in so doing, will allow a tolerance of +/- 0.5% by volume oxygen without a degradation in performance. A fuel-air mixture of 45% to 50% by volume air in hydrogen will give good results.

0056The flow of hydrogen fuel can be increased in order to obtain an initial zero flame temperature of 625°C. This enhanced flame size, in conjunction with an air dilution ratio of 1 part air to 1 part sample, will reduce the oxygen effect to approximately 6% LFL over the range from 0 to 20.9% by volume oxygen in the sample.

0057The 625°C flame with 1:1 air dilution can be zero calibrated using, in place of air, 15% by volume oxygen in nitrogen. This calibration will result in an offset of about 3% LFL, and will alter the analyzer readings to be within +/-3%LFL over the range from 0 to 20.9% by volume oxygen, thereby making the analyzer confirm to the requirement of industry standard accuracy over the entire range of foreseeable oxygen concentrations.

0058The temperature detector can be a thermocouple, a resistance temperature detector (RTD), or any similar device, including an array of thermocouples.

0059The temperature detector can be positioned in the exhaust flow of the sensing flame, or it may be attached directly to the tip of the burner, or within the burner in the vicinity of the sensing flame but not in direct contact with the wall of the burner.

0060The flow rates of fuel and oxidant can be held constant throughout, in order to give a measurement of the flammability of combustible gases without interference from the concentration of oxygen.

0061The flow rates of oxidant can be turned on and off in order to give a measurement of the flammability of combustible gases in the sample and also the concentration of oxygen in the sample.

0062The oxidant pressure regulator may be a mass flow controller that is continuously adjusted by the microprocessor.

0063The burner and detector may be micro-machined in order to optimize the size of the device.