gas sensor calibration

13
161 Chapter 11 Gas Sensor Calibration G as sensors need to be calibrated and periodi- cally checked to ensure sensor accuracy and system integrity. It is important to install station- ary sensors in locations where the calibration can be performed easily. The intervals between calibration can be different from sensor to sensor. Generally, the manufacturer of the sensor will recommend a time interval between calibration. However, it is good gen- eral practice to check the sensor more closely during the first 30 days after installation. During this period, it is possible to observe how well the sensor is adapt- ing to its new environment. Also, factors that were not accounted for in the design of the system might surface and can affect the sensor’s performance. If the sensor functions prop- erly for 30 continuous days, this provides a good de- gree of confidence about the installation. Any possible problems can be identified and corrected during this time. Experience indicates that a sensor surviving 30 days after the initial installation will have a good chance of performing its function for the duration expected. Most problems—such as an inappropriate sensor location, interference from other gases, or the loss of sensitivity—will surface during this time. During the first 30 days, the sensor should be checked weekly. Afterward, a maintenance schedule,

Upload: vankhue

Post on 11-Feb-2017

234 views

Category:

Documents


0 download

TRANSCRIPT

161

Chapter 11 Gas Sensor Calibration

Chapter 11

Gas Sensor Calibration

G as sensors need to be calibrated and periodi-cally checked to ensure sensor accuracy andsystem integrity. It is important to install station-

ary sensors in locations where the calibration can beperformed easily. The intervals between calibrationcan be different from sensor to sensor. Generally, themanufacturer of the sensor will recommend a timeinterval between calibration. However, it is good gen-eral practice to check the sensor more closely duringthe first 30 days after installation. During this period,it is possible to observe how well the sensor is adapt-ing to its new environment.

Also, factors that were not accounted for in thedesign of the system might surface and can affect thesensor’s performance. If the sensor functions prop-erly for 30 continuous days, this provides a good de-gree of confidence about the installation. Any possibleproblems can be identified and corrected during thistime. Experience indicates that a sensor surviving 30days after the initial installation will have a goodchance of performing its function for the durationexpected. Most problems—such as an inappropriatesensor location, interference from other gases, or theloss of sensitivity—will surface during this time.

During the first 30 days, the sensor should bechecked weekly. Afterward, a maintenance schedule,

162

Hazardous Gas Monitors

including calibration intervals, should be established.Normally, a monthly calibration is adequate to ensurethe effectiveness and sensibility of each sensor; thismonthly check will also afford you the opportunity tomaintain the system’s accuracy.

The method and procedure for calibrating thesensors should be established immediately. The cali-bration procedure should be simple, straightforward,and easily executed by regular personnel. Calibra-tion here is simply a safety check, unlike laboratoryanalyzers that require a high degree of accuracy. Forarea air quality and safety gas monitors, the require-ments need to be simple, repeatable, and economi-cal. The procedure should be consistent and trace-able. The calibration will be performed in the fieldwhere sensors are installed so it can occur in any typeenvironment.

Calibration of the gas sensor involves two steps.First the “zero” must be set and then the “span” mustbe calibrated.

Step One: Setting the “Zero” Reading

There is no established standard that defines zeroair. Many analytical procedures, including some spe-cific analyzer procedures such as EPA methods, usepure nitrogen or pure synthetic air to establish thezero point. The reason for this is that bottled nitro-gen and pure synthetic air are readily available. As aresult, it is popularly believed that using bottled nitro-gen or synthetic air is a good method to zero a sensor.

Unfortunately, this is not correct. Normal air con-tains traces of different gases besides nitrogen andoxygen. Also, ambient air normally contains a smallpercentage of water vapor. Therefore, it is much morerealistic and practical to zero the sensor using theair surrounding the sensor when the area is consid-ered to be clean. This reference point can be diffi-cult to establish. Therefore, a good reference point

163

Chapter 11 Gas Sensor Calibration

can be in the area where air is always consideredclean, such as in an office area. This will give a morerealistic representation of the zero point because itwill be representative of the local ambient air condi-tion. The lack of water vapor can cause the zero pointsetting to read lower than in ambient air making thesensor zero appear to drift. This is most noticeablein solid-state sensors and PIDs.

Calibration Methods. Taking all factors such as thetype of sensor and the conditions of the applicationinto consideration, the following are some proposedmethods of calibration:

A. In applications where the ambient air is nor-mally clean, and, based on the operator’s judgmentthat no abnormal condition exists and the instrumentis indicating a close to zero reading, the procedure tozero the sensor can be skipped. When in doubt, use aplastic bag to get a sample of what is considered to be“clean air” in the facility and expose it to the sensorfor a few minutes. This is a very quick and easy proce-dure. It is also a very effective way to differentiate areal alarm from a false alarm.

B. Compressed air has the advantage that it is easyto regulate and can be carried around in a bottle. Also,in many facilities, shop air is available throughout theplant, making it very accessible and convenient. How-ever, most shop air contains small concentrations ofhydrocarbons, carbon monoxide, carbon dioxide, andpossibly other interference gases. Also, the air is typi-cally very low in humidity. A solution to this is that theair can be filtered through activated charcoal to re-move most of the unwanted gases and water vapor canbe added into the air using a humidifier in the sam-pling system. After this conditioning, the air can beused to calibrate most types of sensors. However, it isimportant to note that carbon monoxide is not re-moved by charcoal filters.

It is therefore imperative to make sure that the

N2

O2

H2O

CO2 & others

Ambient air is the best zero air.

164

Hazardous Gas Monitors

CO concentration in the shop air is the same as in theambient air. Furthermore, a soda ash filter should beused to remove carbon dioxide. This is also a very goodway to zero carbon dioxide sensors since placing a sodaash filter in-line with the sampling system will removeall carbon dioxide, thus providing an easily obtain-able zero baseline.

Although synthetic air is usually very pure, it can-not be used with solid-state sensors or PID sensorsbecause these sensors require some water vapor in thesample stream. A simple solution to this problem is toadd a wet tissue paper in the sample line. This acts asa humidifier in the sample stream and providesenough water vapor for the sensor to read properly.Another option is to use a Nafion tube, which is de-scribed thoroughly in Chapter 10, “Sampling Systemsand Designs.” Figure 1 illustrates this concept.

Fig. 1 Adding Moisture to Calibration Gases

Step Two: Span Calibration

The span calibration can be quite easy or it can bevery complicated and expensive, depending on the gastype and concentration range. In principle, to achievethe best accuracy, a mixture of the target gas balanced in thebackground environmental air is the best calibration gas.However, although this can be done, it usually requiresthat the operators be more skilled than would usuallybe required. In practice, most calibration gases are pur-chased from commercial suppliers. The following sec-tion describes a few methods of span calibration.

A. Premixed Calibration Gas

Calibration Gas

Regulator

NafionDryer Tube or

Wetting MaterialTo SensorFlow

Meter

165

Chapter 11 Gas Sensor Calibration

This is the preferred and most popular way to cali-brate gas sensors. Premixed gas mixtures are com-pressed and stored under pressure in a gas bottle. Thebottles are available in many sizes but most field cali-brators employ smaller, lightweight bottles. Thesesmall portable bottles come in two different catego-ries: a low-pressure and a high-pressure version.

The low-pressure bottles are thin-walled, light-weight bottles that are usually nonreturnable and dis-posable. High-pressure bottles are designed to bottlepure hazardous chemicals. For calibration gases, thesebottles are normally made of thick-walled aluminumwhich has a service pressure of 2000 psi.

To get this highly pressurized gas out of the bottlein order to calibrate the sensor, a regulator assemblyis needed. This assembly consists of a pressure regu-lator, a pressure gauge, and an orifice flow restrictor.The orifice flow restrictor is a fitting with a hairlinehole that allows a constant air flow at a given pres-sure difference. In operation, the high pressure fromthe bottle is reduced to a lower pressure of only a fewpsi, which provides a constant air flow through theorifice. Flow rates between 600-1000 cc/min are most common. Models can befitted with an adjustable pressure regu-lator so that the flow rate can be adjustedaccordingly. Figure 2 illustrates typicalmodels of high- and low-pressure bottleassembly.

Many gases can be premixed withair and stored under pressure, but somegases can only be mixed in inert gasbackgrounds, such as nitrogen. Somemixtures can only be stored in bottlesthat are specially treated or condi-tioned. Each type of mixture will havea different amount of time before it ex-pires or before it can no longer be used.

Low-Pressure Assembly High-Pressure Assembly

Fig. 2 Calibration Gas Bottles

166

Hazardous Gas Monitors

Detailed information about storage and shelf life canbe obtained from the manufacturer. Generally, highvapor pressure gases with low reactivity, such as meth-ane, carbon monoxide and carbon dioxide, can bemixed with air and stored under high pressure. Lowvapor pressure gases, such as liquid hydrocarbon sol-vents, can only be mixed with air and stored underlow pressure. Most highly reactive chemicals are mixedwith a nitrogen background. With certain sensors, suchas solid-state sensors, whether the mixture of the gasis in the air or in the nitrogen background will dra-matically affect the sensor reading.

During calibration, some sensors may need mois-ture to get a proper reading. Moisture can be added byfollowing the same procedure described in Step 1 forzeroing the sensor.

To estimate the volume of a pressurized gas in acylinder, take the total pressure (P) divided by the at-mospheric pressure (Pa) and multiply this ratio by thevolume of the cylinder:

Vmix = V . (P/Pa)

where

Vmix = the volume of the gas mixtureV = the volume of the cylinderP = the pressure in the cylinderPa = the atmospheric pressure

For example, let’s say a given lecture bottle has a 440cc volume (V). Assume the bottle has a 1200 psi pres-sure. The estimated volume of the premixed gas at atmo-spheric pressure is: (440 cc) x (1200/14.7) = 35,918 cc.

If the flow rate of the calibration gas is 1000 cc perminute and it takes approximately one minute per sen-sor to calibrate, a single cylinder can be used to cali-brate approximately 30 times.

167

Chapter 11 Gas Sensor Calibration

B. Permeation DevicesA permeation device is a sealed container that con-

tains chemicals in liquid and vapor phase equilib-rium. The gas molecules are either permeatedthrough the permeable container wall or throughthe end cap. The rate in which the gas moleculespermeate depends on the permeability of the mate-rial and temperature. The rate of permeation is con-stant over long periods of time. At a known rate ofpermeation at a given temperature, a constant flowrate of air mixed with the permeated chemicals formsa constant stream of calibration gas. A calibrator withconstant temperature and flow regulation is needed.However, the permeation tube continuously emitschemicals at a constant rate thus creating a storageand safety problem. Also, the rate of permeation fora given gas of interest can be too high or too low fora given application. For example, high vapor pres-sure gases permeate too quickly while very low vaporpressure chemicals have a permeation rate that is toolow to be of any use.

Permeation devices find most of their use in labo-ratories and in applications using analytical analyzers.For gas monitoring applications, the concentrationsneeded to calibrate the sensor are typically too highfor the permeation device. Therefore, they have beenfound to be of limited use.

C. Cross CalibrationCross calibration takes advantage of the fact that

every sensor is subject to interference by other gases.For example, for a sensor calibrated to 100% LEL hex-ane, it is usually much easier to use 50% LEL meth-ane gas to calibrate the sensor instead of using an ac-tual hexane mixture. This is because hexane is a liq-uid at room temperature and it has a low vapor pres-sure. Therefore, it is more difficult to make an accu-rate mixture and to keep it under high pressure.

Examples of Permeation Tubes

168

Hazardous Gas Monitors

On the other hand, methane has a very high va-por pressure and is very stable. Furthermore, it canbe mixed with air and still be kept under high pres-sure. It can be used for many more calibrations than ahexane mixture in the same size bottle and it has along shelf life. A 50% LEL methane mixture is alsoreadily available. Therefore, it is common practice formanufacturers of combustible gas instruments to rec-ommend the use of methane as a substitute to cali-brate for other gases.

There are two ways to accomplished this task. Thefirst method is to calibrate the instrument to meth-ane while other gas readings are obtained by multi-plying the methane reading by response factors thatare included in the manual. This is commonly donewith catalytic sensors. Catalytic sensors have a linearoutput and therefore the use of this response factor isapplicable to the full-scale range. For example, pen-tane has an output of only half that of methane gaswhen the sensor is calibrated to methane. Therefore,it has a response factor of 0.5. So, if the instrument iscalibrated to methane but is used to measure pentane,the reading is multiplied by 0.5 to obtain the pentanereading.

The second method is to still use methane as thecalibration gas, but double the value of the reading ofthe calibration. For instance, use 50% LEL methanecalibration gas and calibrate with this as 100% LELpentane. After the calibration, the instrument directlyindicates the pentane gas concentration although itwas calibrated using methane gas.

Many low-range toxic gas sensors can be calibratedusing cross gas calibration. Also, with infrared instru-ments, any gas within the same wavelength of absorp-tion can be used for cross calibration. The advantageof cross calibration is that it allows the sensor to becalibrated with a gas and range that is easier to obtainand handle.

169

Chapter 11 Gas Sensor Calibration

However, there are some problems with using crosscalibration. One is that the response factors for eachsensor can be different as it is generally impossible tomake most sensors exactly alike. For example, in cata-lytic sensors, the heater voltage has to be as specifiedin the manual; otherwise, the response factor will notbe applicable. The response characteristics will varywith different heater voltage settings. Therefore, it isa good practice to periodically check the calibrationof the sensor with the actual target gas.

Mixtures of stable noncombustible and nontoxicgases with various concentrations are available frommany supply sources. Check with the instrumentmanufacturer for more detailed information.

D. Gas MixingNot all calibration gases are available. Even if they

are available, it is very possible that they would not beavailable in the right concentration or in the properbackground mixture. However, many mixtures areavailable for some process uses which can be dilutedto use in calibration of gas monitors in lower concen-tration ranges. For example, 50% LEL methane hasa concentration of 2.5% or 25,000 ppm. To make a20% LEL mixture having a volume of 2000 cc, thefollowing formula can be used:

Vb =C

. V, Va =C – Cb . V

Cb C

and

Va = V – Vb

whereCb = concentration in the bottle, 50% in this

caseC = new concentration, 20% in this caseV = total final volume, 2000 cc in this caseVb = volume of mixtureVa = volume of air or other dilutant

170

Hazardous Gas Monitors

Vb = 20/50 . 2000 = 800 cc

Va = 2000 – 800 = 1200 cc

The final mixture would be made by taking 800 ccof the calibration gas and mixing it with 1200 cc of airto make the mixture equal to 20% LEL.

Another example is to dilute this 25000 ppm ofmethane calibration gas to make a 100 ppm of mix-ture.

Vb = 100/25000 . 2000 = 8 cc

therefore

Va = 2000 – 8 = 1992 cc

By mixing 8 cc of calibration gas into 1992 cc ofair, 2000 cubic centimeters of 100 ppm gas mixture isobtained.

Some Calibration Tools

To perform the above procedure, the followingtools are needed:

1. Syringe and Needle: This is themost inexpensive way to accuratelymeasure the amount of gas. A dis-posable medical syringe with a largegage needle is most practical but

there are few syringes with more than one hundredcubic centimeter volume. Hence, large volume mea-surements can be troublesome. However, it is easy tomake a syringe using any standard size pipe havingabout a 2-inch diameter. It provides an easy and con-venient means to make a mixture on a regular basis.For very small volume measurements, there are mi-cro syringes that are readily available in chemicalsupply catalogues.

2. Calibration Bag: Most of the materials used infood packaging or storage are quite inert; otherwise,food would be contaminated with odor. Therefore,

A 2”-Diameter, 1000 cc Syringe

Standard Medical Syringes:1 cc, 10 cc, and 50 cc

171

Chapter 11 Gas Sensor Calibration

food storage bags can be used to hold most chemicalsas long as they are used for relatively short durations.This is an important point to keep in mind since gasmolecules will eventually diffuse through the manythin layers of a plastic bag. For example, potato chipscan stay fresh in their original bag for long periods oftime because the bag material is less permeable bygas molecules than normal food storage bags. This isdemonstrated by the fact that when the potato chipsare transferred into a tightly sealed food containerbag, they will lose their crispness in a very short time.There are also many commercially available samplingbags on the market. One common example is a Tedlarbag. It is made from polyvinyl fluoride and has lowabsorption of gas molecules. However, this type of bagis still permeable, so a heavy gauge material will beneeded if permeability is a major concern. Samplingbags normally come with a valve and a septum that isused as an injection port.

Pressure Formula

Earlier, we described preparing a mixture basedon a volume relationship. Based on the ideal gaslaw, the same volume formula can be used as a pres-sure formula. As an illustration, take an 800 psi mix-ture of 50%LEL methane with a 1200 psi mixtureof air. This will result in a 2000 psi mixture of20%LEL methane.

Preparing gas mixtures can be a very difficult task.It is best to consult with the instrument manufac-turer regarding the best method of calibration andavailability of gas mixtures.

Following are some examples of how gas mix-tures can be made:

For ppm gas mixtures:

Cppm = Vc /(Vc + Vd) . 106 ppm

where Vc is target gas volume and Vd is the

A 5-liter Sampling Bag

1000 cc Calibration Cans

Mixing Calibration Gas

172

Hazardous Gas Monitors

dilutant volume. For example, what is the concentra-tion in parts per million when 1 cc of CO is added toa 1000 cc container?

Cppm = 1/1000 . 106 = 1000 ppm

For % range gas mixtures:

C% = Vc /(Vc + Vd) . 102 %

The Vc term in the denominator can be insignifi-cant in low ppm mixtures.

Calibrating Liquid Chemical Mixtures. To make acalibration mixture for liquid chemicals, a known vol-ume of liquid is vaporized in a known volume ofdilutant air. The ideal gas law states that one grammole of molecules will occupy 24,500 cc of volume at25 degree centigrade and at 760 mm of mercury orsea level atmospheric pressure. This temperature andpressure is also called the standard condition. At stan-dard conditions, the equation is:

Cppm = 24.5 . 109 . (V x D)/(Va . M)

where V = volume of liquid, D = density of the liq-uid, which is the same as the specific gravity, Va = thevolume of the dilutant air, and M = the molecularweight of the liquid.

Since it is easier to measure the liquid using a mi-cro syringe, the equation then becomes

V = Cppm . Va . M/(24.5 . 109 . D)

where all units are in milliliters, cubic centimeters,and grams.

For example, benzene has M = 78.1 g and D =0.88 g/cc. What is the amount of benzene needed tomake a 1000 ppm mixture in a 2000 cc bottle?

V = 1000 . 2000 . 78.1/(24.5 . 109 . 0.88)

which yields

V = 7.2 . 10-3 = 0.0072 cc = 7.2 microliters

Many containers are sized by gallons.Therefore, it is useful to know thatone gallon is equal to 3785 cc.

173

Chapter 11 Gas Sensor Calibration

In air pollution, industrial hygiene, and medicaltoxicology work, the commonly used unit of concen-tration is milligram per cubic meter. The followingequation expresses this relation, again assuming stan-dard conditions.

Cppm = C . 24.5/M

where

C = mg/m3

M = molecular weight

In conclusion, for the calibration of gas monitors,accuracy is not extremely important because these arenot analytical devices or systems. However, it is most im-portant to keep the calibration methods standardized andeasily traceable. If procedures are standardized, datacan be normalized at a later date if necessary.

Ventilation Hood for Calibration Procedures