Understanding Modern Infrared Pyrometers for Demanding Steel Mill Applications By Tom Larrick , Williamson Corporation, Concord MA , USA Advanced innovations and modern digital electronics have revolutionized infrared pyrometers in recent years. The steel industry has directly benefitted from the introduction of application-specific infrared pyrometers to produce reliable temperature readings under demanding mill conditions. This paper briefly describes recent single, dual, and multi-wavelength pyrometer developments pertaining to critical steel mill measurements. Included are traditional steel mill issues such as compensation for emissivity, water, scale, dust, dirty optics and heat, as well as more topical issues such as the measurement of dual-phase steel, stainless steel, coated steel, and cold-rolled strip. Most modern infrared pyrometers are highly accurate, precise and stable when used in a clean and controlled calibration laboratory; yet, inconsistent and unreliable performance is frequently obtained when general-purpose sensors are applied throughout the steel mill. Emissivity variation, dirty optics, scale, steam, water spray, oil, flames, combustion byproducts, hot background reflections, mechanical abuse and high ambient temperature exposure represent a sampling of the significant interference sources encountered by steel mill users. Laborious maintenance procedures, sophisticated installation accessories, or elaborate compensation routines are often developed in an effort to produce the required degree of accuracy for precise process control, but these techniques usually fall short of perfection. As a result, process consistency, product quality, and plant efficiency suffer; cumbersome maintenance demands persist, and user confidence falters. These problematic application issues unique to the steel industry are best addressed through the selection and use of infrared pyrometers designed to operate with these common measurement conditions in mind. Infrared pyrometers are popular in the metals industry because of the ability to measure product temperature from a distance. These sensors are ideal for measuring hot, moving targets, or for measuring in a hostile environment where traditional contact temperature sensors are not appropriate. These devices do not measure temperature directly; instead, they infer a temperature value based upon the infrared energy emitted by the target of interest. The infrared energy is measured and converted into an electrical signal by an internal infrared detector. The infrared pyrometer calculates a temperature value by correlating the measured infrared energy to a calibrated temperature value. The resultant temperature measurement is accurate so long as the assumptions inherent in the pyrometer design accurately represent operating conditions encountered by the sensor. An error in temperature measurement will occur when application conditions exist that are not accounted for in the pyrometer design. Every infrared pyrometer may be categorized into one of three sensor types: single-wavelength (one-color), ratio (dual-wavelength or two-color), and multi-wavelength. See Figure 1. Each type of pyrometer plays an important role for various applications throughout the iron and steel industry, and each makes different assumptions about the measurement conditions. Single-wavelength, or one-color, pyrometers are accurate when the emissivity is constant and known, when the optical path between the pyrometer and target remains clear, and when reflected background influences are negligible. Ratio pyrometers, also known as dual-wavelength or two-color sensors are used when the emissivity is variable or unknown, or when the optical path is partially obstructed. Just as the single-wavelength pyrometer assumes that the emissivity of the measured material is constant and known, the ratio pyrometer assumes that the ratio of emissivity values at the two measured wavelengths is constant and known. Multi-wavelength pyrometers, are used when the emissive character of the measurement does not permit the use of traditional single-wavelength or ratio sensors. This type of pyrometer assumes that the emissive

Page 1

character of the measured material may be characterized by an application-specific algorithm. Each type of pyrometer has its own strengths and weaknesses; however, within each type of infrared pyrometer, performance may be optimized and sensitivity to interference may be reduced through thoughtful pyrometer design. Sensor wavelength is perhaps the most important consideration in thoughtful pyrometer design. As Figure 2 illustrates, infrared pyrometers filtered at relatively short wavelengths are considerably more sensitive to changes in temperature than are sensors filtered at longer wavelengths. The sensitivity of single-wavelength pyrometers to a 10% change in emissivity or a 10% optical obstruction (10% reduction of infrared energy) is shown in Figure 3. This figure shows that measurement errors are significantly smaller when using a short-wavelength singlewavelength pyrometer compared with a long-wavelength pyrometer. Temperature errors also decrease for both long-wavelength and short-wavelength pyrometers when measuring at low temperatures. Indeed, the improved accuracy of a short-wavelength sensor is understated by the theoretical values depicted in Figure 3. This is because the emissivity value of a metal surface is almost always significantly higher at shorter wavelengths than it is at longer wavelengths (Figure 4), making any given absolute change in emissivity a smaller percentage change. For example, a change in emissivity from 0.30 to 0.31 is about a three percent change. On the other hand, a change from 0.10 to 0.11 is a ten percent change. Particularly for low emissivity materials, a short-wavelength pyrometer is typically 4 to 16 times more accurate than a long-wavelength pyrometer due to the inherent lower sensitivity to and smaller percentage variation from changes in emissivity. Recent advances in short-wavelength pyrometer technology allow these short-wavelength sensors to rival the broad temperature spans historically available only when using a long-wavelength pyrometer. Many steel mill applications requiring broad temperature spans, such as refractory preheating and reheat furnace entry measurements, may now take advantage of short-wavelength technology to dramatically reduce measurement errors. The fact that a short wavelength sensor is more accurate when confronted with optical obstruction and emissivity variation is generally known and understood by experienced steel industry users, but the importance of wavelength selection does not end with this fundamental rule. General purpose short-wavelength sensors filtered over a waveband of 0.7 to 1.1 microns, 1.0 to 1.6 microns, or 2.0-2.7 microns are commonly available and routinely employed by the steel industry. As illustrated by Figure 5, these popular wavebands are selected by most pyrometer manufacturers because they correspond to the range of highest response for the semiconductor detectors used to convert the infrared energy into an electrical signal. Unfortunately, none of these general purpose wavebands is appropriate for applications where steam, flames, or combustion gasses are present. Just as fog prevents humans from seeing clearly through moisture-laden air, pyrometers filtered over these common wavebands cannot clearly see through these typical interference sources resulting in a measurement error whenever they are encountered. Careful wavelength selection is therefore critical for accurate measurement throughout the mill where these common interferences are present. These interferences can be found at the hot mill (especially at the quench area and coiler, and especially for dual-phase steel which must be cooled faster and lower than traditional alloys), at the coke conveyor, in the ladle and tundish preheat stations, in the reheat furnace and heat treat furnaces, for annealing lines, and even for hot metal detectors when water or steam is encountered. Because temperature measurements are commonly made within gas-fired furnaces, interference from combustion gas is a common and often undiagnosed cause of temperature measurement error. Figure 6 demonstrates that optimal wavelength bands exist throughout the infrared spectrum where water vapor, carbon dioxide, and carbon monoxide gasses are highly transparent. The selection of a pyrometer filtered within a wavelength band of high optical transmission is essential for singlewavelength, ratio, and multi-wavelength pyrometers alike whenever water, steam, flames or products of combustion are present. For example, Figure 7 shows the optical transmission of infrared energy through two different thickness of water. A ratio sensor can tolerate an optical obstruction so long as both wavelengths are obstructed equally. Two wavebands selected at wavelengths in the 0.7 and 0.8 micron range will view through the water without interference, sensors filtered at 0.8 and 0.9 microns will indicate some minimal interference, and sensors filtered at 0.7 and 1.1 micron will indicate a relatively large interference. This is because the water absorbs much more infrared energy at wavelengths beyond one micron than at wavelengths shorter than one micron.

Page 2

Additional wavelength considerations exist when selecting the most appropriate ratio sensor. Like the legs of a table, a ratio sensor is unstable when the wavelength pairs are too close together. Most ratio sensors use a bi-level two-element detector to separate the measured wavelength into two wavebands. The responsivity curve for one such detector is shown in Figure 8. The top element is opaque at and sensitive to shorter wavelengths, and is translucent at longer wavelengths; thereby allowing some infrared energy at the longer waveband to pass through to the lower detector element. The lower detector element is sensitive to infrared energy at this longer wavelength range. This relatively straightforward approach effectively separates the infrared energy into two wavebands, as is required for a ratio pyrometer. However, the resultant wavelength bands overlap each other with absolutely no wavelength separation, leaving it relatively unstable to potential optical obstructions and spectral emissivity (eslope) variation. For ratio pyrometers, robust sensor performance in the steel mill demands the selection of wavelengths able to view through common interference sources and the need for separation between wavelengths. Dual-wavelength pyrometers designed with these issues in mind can view through water, steam, and combustion byproducts; are as much as 20 times less sensitive to scale formation; and can tolerate 20 to 60 times more optical obstruction (dirty lens, for example) as compared to traditional, general-purpose ratio sensors with overlapping wavebands (twocolor pyrometers). Through thoughtful wavelength selection, these powerful and robust dual-wavelength pyrometers may be used with confidence where other pyrometers are not able to operate:  Where water is encountered in the caster containment zone, and for underside measurements at the plate mill roughing stand.  Where severe optical obstruction is commonly encountered at the coke guide and when measuring molten iron stream temperature at the blast furnace.  Where heavy scale is encountered at casters, rolling mills, and some heat treating furnaces.  Where significant temperature gradients exists at the sinter and briquette plants, and at the welders.  Where there is a need to view through combustion gasses in the soaking zone of heat treating furnaces. While significant advancements have been made in both single-wavelength and ratio sensor technology, perhaps the most significant development is the multi-wavelength infrared pyrometer. Long considered problematic and difficult to work with, these sophisticated pyrometers have matured to become highly accurate, reliable, easy to use, and have gained wide-spread acceptance for specific applications within the steel industry. Each multiwavelength pyrometer is specifically designed to account for the unique emissivity character associated with a particular problematic measurement, material, or application. Multi-wavelength pyrometers have been developed for the measurement of many non-greybody materials (those for which the emissivity value changes with wavelength), but those models most applicable for the steel mill industry are those designed for use on cold-rolled steels, and coated steel (zinc, aluminum, magnesium, etc.). Figure 9 compares the temperature error of the multi-wavelength infrared pyrometer with the single-wavelength and ratio models for the measurement of steel strip. The y-axis indicates temperature error, while the x-axis indicates the steel strip emissivity. Note that a single-wavelength sensor is only accurate when the strip emissivity value is equal to the emissivity parameter setting of the sensor, in this case 0.3. If the strip emissivity is lower than the emissivity setting, then the pyrometer will read low. If the strip emissivity is higher than the emissivity setting, then the pyrometer will read high. A good example of this is illustrated in Figure 10, which shows the temperature error associated with a single-wavelength pyrometer with an emissivity setting fixed at 0.45, compared with actual emissivity values. This level of error with a single-wavelength pyrometer is unacceptable to be used as a control parameter for a heat treating process, and is fixable with either a dual- or multi-wavelength pyrometer. Also note that the ratio sensor will measure the correct temperature value so long as the strip emissivity value is higher than about 0.6, but will read a value that is too high for lower emissivity values. When using a ratio pyrometer at emissivity values above 0.6, the strip is sufficiently coarse at the microscopic level so that the emissivity is equal at both measured wavelengths. However, as the strip emissivity drops below the 0.6 value, the strip becomes smoother at the microscopic level, and the emissivity starts to differ at the two measured wavelengths. The lower the strip emissivity, the greater the difference in emissivity values at the two measured wavelengths, and the greater the error indicated by the ratio pyrometer. A multi-wavelength

Page 3

infrared pyrometer is able to characterize this emissivity variation across the measured wavelengths and to make the appropriate compensation. Hot rolled steel strip generally has a relatively high emissivity value, allowing a ratio sensor to produce a highly accurate temperature reading; however, cold rolled steel strip generally has a low emissivity value. The low emissivity nature of cold rolled strip indicates the necessity to use a multi-wavelength infrared pyrometer, particularly when measuring at the relatively high temperatures associated with the annealing process or when heating prior to a hot-dip zinc-coating. When measuring cold rolled steel strip at the relatively low temperatures encountered at the cold mill, temper mill, and coating line preheat locations, a short-wavelength singlewavelength pyrometer usually provides the required degree of accuracy. It is not uncommon to use a combination of multi-wavelength and single-wavelength pyrometers on the same line when temperature values range from low to high. Because the multi-wavelength pyrometer measures both the temperature and the emissivity value, it is possible to feed-forward the real-time emissivity measurement to the single-wavelength sensor for more precise measurements. Multi-wavelength pyrometers are commonly used in the annealing furnaces for all steels, including mild alloys, stainless alloys, high-strength alloys, high-temperature alloys, and electrical alloys. Multiwavelength sensors are also used to provide an accurate temperature measurement throughout the galvanneal process, for spangle size control of galvalume, and for pick-up prevention of all hot-dipped zinc-coated products as they pass the up-leg turn-roll. Even the most accurate infrared pyrometer is of no use to the steel industry if it is not mechanically durable enough to survive the hostile mill environment. Fiber optic cables are commonly used to align the pyrometer to the measured target when there is limited space or under hostile operating conditions. Traditional fiber optic cables do not offer a high level of protection against mechanical and thermal damage, and fiber cable damage is a common maintenance issue. Heavy duty fiber cables with rugged stainless steel armor and multiple layers of thermal insulation, such as that shown in Figure 11, provide a degree of protection more in keeping with the rigorous demands of the steel industry. Rugged fiber optic cables such as this may be used with confidence for measurements throughout the mill, such as at vertical ladle preheat stations, continuous casters, coke guides, blast furnaces, and underneath the hot mill. Some sophisticated infrared pyrometers offer advanced signal conditioning features that are of significant value to the steel industry. For those cases when an intermittent interference in the sensor reading cannot be avoided, some modern ratio and multi-wavelength infrared pyrometers measure emissivity and infrared energy values in addition to the calculated temperature value. When unavoidable interference sources such as intermittent flames or heavy sparklers enter the sensor’s view, one or more of these extra measured parameters often jumps out of the normal range associated with a valid measurement condition (Figure 12). By establishing the limits associated with normal operating conditions, the sensor may be configured to ignore any readings taken when the interference source is present. This feature is particularly useful when measuring molten iron and steel stream temperatures at the blast furnace or melt shop, and is also used to ignore fallen or flying hot scale at the hot mill. Advanced signal conditioning features such as this compliment traditional signal conditioning functions to produce stable and reliable temperature measurements under a wide range of challenging plant conditions. The infrared pyrometer industry has matured over the past decade, so that the steel industry no longer needs to accept mediocre performance, and can instead insist on precision and reliability. Application-specific pyrometers will provide more accurate temperature measurement and will provide years of low-maintenance service, as compared with general purpose infrared pyrometers. Thoughtful attention to wavelength selection will improve sensor accuracy and performance for applications where steam, flames, combustion gasses, water, scale, or significant temperature gradients are potential sources of interference. Thoughtful wavelength selection will also eliminate the need for extravagant installation accessories and error compensation schemes. Accurate temperature measurement and control is crucial throughout many parts of the steel mill as Figure 13 illustrates.

Page 4

The following pyrometer technologies are recommended for the most accurate and reliable temperature measurement for these specific applications:      



  

Stove Dome: Dual-wavelength pyrometer with fiber optics to look through dirty windows and environment to measure the refractory temperature. Ladle and Tundish Preheaters: Single-wavelength pyrometer that looks through flames with fiber optics and ArmorGuard to prevent refractory overheating and save on energy costs. Caster: Dual-wavelength pyrometer with fiber optics and ArmorGuard for slabs and strands. This design looks through water and spray to control the containment section for safety and quality (cracking and nonuniformity) issues. Hot Rolling Mill: Dual-wavelength pyrometer with visual aiming for top side and fiber optics for bottom side that looks through water and spray to profile product temperature from furnace exit to coiler. Annealing: Multi-wavelength pyrometer for non-greybody steel strip. Finishing Mill Coating: Multi-wavelength pyrometer for metal coatings (e.g., Zinc, Tin, Al, Mg, etc.) at high temperatures and short-wavelength, single-wavelength for the low temperatures. Galvanneal stock needs close temperature control. Paint coating requires selective narrow band filtering to measure the thin paint coating. Hot Mill Reheat Furnace: Dual-wavelength pyrometer for the soaking zone when there is a flame free path between the sensor and the slab. Single-wavelength pyrometer for coal or oil-fired furnaces, when viewing through flames, and installation in the heating zones. Both dual and single-wavelength technologies require careful wavelength selection. Hot Mill Intermediate Quench: Single-wavelength pyrometer with a carefully selected wavelength able to view through heavy steam and moderate levels of water. Blast Furnace Iron Stream: Dual-wavelength pyrometers compensate for dirty optics and self-align to the molten iron stream. Thoughtful wavelength selection eliminates interference from combustion gasses, flames, and smoke. Blast Furnace - Tuyere: Dual-wavelength pyrometers compensate for dirty optics and self-align to the blowhole. Thoughtful wavelength selection eliminates interference from process gasses, flames, and smoke.

Illustrations Figure 1 = Three Types of Infrared Pyrometers (Single-Wavelength, Ratio, and Multi-Wavelength) Figure 2 = Infrared Energy vs. Temperature vs. Wavelength Curve Figure 3 = Error for a 10% Change in Emissivity or 10% Optical Obstruction (Single-Wavelength Pyrometers) Figure 4 = Emissivity vs. Wavelength for Low-Emissivity Materials Figure 5 = General-Purpose Wavelength vs. Thoughtful Wavelength Selection to Tolerate Common Industrial Interferences Figure 6 = Atmospheric Absorption Bands Showing H2O and CO2 Figure 7 = Water Transmission Showing Popular Ratio Wavelengths Figure 8 = Two-Color Detector Responsivity Curve Figure 9 = Single, Dual, and Multi-Wavelength Measurement of Steel Strip Figure 10 = Single-Wavelength Errors on Annealing Line Figure 11 = ArmorGuard Fiber Optic Cable Figure 12 = ESP Filtering Figure 13 = Steel Mill Application Overview References D.P DeWitt, Theory and Practices of Radiation Thermometry, Copyright 1988 W.R. Barron Jr, Tackling the Tough Processes, Process Heating, January 2004 page 39 G.G. Gubareff, Thermal Radiation Properties Survey, Honeywell Res. Center 1960, page 269

Page 5

Three Types of Infrared Pyrometers All infrared pyrometers measure infrared energy. Because infrared energy is emitted by an object as a function of temperature and emissivity, all infrared pyrometers calculate a temperature value based upon an assumption about the emissivity character of the measured material.

Single-Wavelength (One-Color) Pyrometers Assumption: Emissivity is constant and known. Single-wavelength pyrometers measure the amplitude of infrared energy collected over a specific wavelength span. These sensors provide an average temperature, and are affected by changes in emissivity, dirty optics, and other optical obstructions. Stray infrared energy from background sources will affect the sensor reading if the energy is significant. Sensitivity to emissivity variation, optical obstructions, and background energy varies with wavelength. Wavelength selection can significantly impact the ability to view through certain intervening media, such as steam, flames, or combustion gasses.

Ratio (Dual-Wavelength or Two-Color Pyrometers) Assumption: The ratio of emissivity at wavelengths A & B is reasonably constant and known (the definition of a greybody material). Ratio pyrometers measure the ratio of energy collected at two adjacent wavelengths. The ratio value is not affected by so called “grey” obstructions (those that obstruct both wavelengths equally). As such, these sensors are able to correct for emissivity variations, and to view through smoke, dust, and most other optical obstructions. This sensor tends to provide a very heavily weighted average towards only the hottest temperature viewed. Compared to a single-wavelength pyrometer, stray infrared energy from a hotter background will affect the sensor reading more, and from a cooler background, less. Sensitivity to variations in the ratio of emissivity at wavelengths A and B (emissivity slope, or e-slope) varies by wavelength pair. Similarly, the sensitivity to some types of intervening media such as water, steam or combustion byproducts will vary with wavelength selection.

Multi-Wavelength Pyrometer Assumption: The effective target emissivity (or other emissivity-related parameter such as e-slope) is related to some measurable parameter (such as measured signal strength or signal dilution) with a one-to-one correlation. Multi-wavelength sensors measure infrared energy at two or more wavelengths, and calculate the target temperature based upon an application-specific model. This model is formed using a one-to-one correlation between a corrective emissivity adjustment (or some other emissivity-related parameter, such as e-slope) and a meaningful measured parameter, such as signal strength or signal dilution, or a function of the two. Multiwavelength algorithms may include additional application information, such as background temperature or a measured reflectivity value, and they may involve multiple iterations of calculations. These sensors are used for materials and applications where traditional models have proven to be inappropriate.

Figure 1 = Three Types of Infrared Pyrometer (Single-Wavelength, Ratio, and Multi-Wavelength) Page 6

Figure 2 = Infrared Energy vs. Temperature vs. Wavelength Curve

Page 7

Figure 3 = Error for a 10% Change in Emissivity or a 10% Optical Obstruction (Single-Wavelength Pyrometers)

Page 8

Figure 4 = Emissivity vs. Wavelength for Low-Emissivity Materials

Page 9

Figure 5 = General-Purpose Wavelength vs. Thoughtful Wavelength Selection to Tolerate Common Industrial Interferences

Page 10

Figure 6 = Atmospheric Absorption Bands Showing H2O and CO2 Atmospheric transmittance at sea level over a 0.3 km path. Atmospheric absorption and emission is a major factor in the selection of the wavelength and bandwidth for use in radiation thermometry. Absorption varies with gas concentration and optical pathlength. Adapted from H.W. Yates and J.H. Taylor, Infrared Transmission of the Atmosphere. [8]

Page 11

Figure 7 = Water Transmission Showing Popular Ratio Wavelengths

Page 12

Figure 8 = Two-Color Detector Responsivity Curve

Page 13

Figure 9 = Single, Dual, and Multi-Wavelength Measurement of Steel Strip.

Page 14

Analysis of Temperature Variance Sample Coils from Dec 4th and 6th 30

2

25

1.9

Ircon lessLess Williamson Single Multi Actual Coil Emissivity Actual Coil Emissivity Single Emissivity Setting Ircon Emissivity Setting

20

1.8 1.7

15

1.6 10 1.5 1.4

0

1.3

-5

1.2

-10

1.1

-15

1 0.9

-20

0.8

-25

0.7 -30 0.6 -35

0.5

-40

0.4

-45

0.3

-50

0.2 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

Coil

Figure 10 = Typical Single-Wavelength Errors on Annealing Line

Page 15

Emissivity

Temperatre Variance (°F)

5

Figure 11 = ArmorGuard Fiber Optic Cable

Figure 12 = ESP Filtering

Page 16

Figure 13 = Steel Mill Application Overview Page 17