The development of industrial temperature measurement technology has transcended the stage of predictive maintenance, becoming a major sensing technology. Such process industry applications include the production of glass products and plastic packaging films. As infrared thermal imaging applications for predictive equipment maintenance become increasingly prevalent, factory engineers are increasingly aware that the superior performance of key sensors can ensure control systems remain under their control. Infrared thermal imaging technology, or thermal imaging, maps the surface temperature of an object by monitoring the infrared radiation it emits. An infrared thermal imager assigns different colors to different temperatures, allowing the naked eye to identify very narrow temperature ranges. All thermal imagers use "false colors" in their images because the actual colors are located in the infrared range, which is invisible to the naked eye. More importantly, such devices also have the ability to analyze images and extract precise control information. The Development Trend of Thermal Imaging Technology Historically, the development of thermal imaging technology has relied on adapting itself to new application areas, consolidating its position, and then expanding into other application areas. Its earliest industrial applications began in the 1980s for failure analysis of electronic component assemblies. Engineers analyze thermal images of circuit boards that fail electrical inspections in hopes of modifying designs or processes to improve production. [align=center]Figure 1: A thermal image of a heater shows temperature variations; blue areas are the hottest, and red areas are the coldest.[/align] Some designers also use thermal imaging to analyze heat flow on prototype boards. In the absence of computer thermal flow simulation (discussed later), thermal imaging provides the fastest and cheapest tool for heat source control in electronic assemblies. By the 1990s, equipment engineers began developing predictive maintenance procedures combined with routine temperature monitoring. They periodically collected temperature parameters from equipment such as electric motors, gearboxes, transformers, and other power devices. Initial mechanical failures, such as bearing wear, insulation breakdown, and contact corrosion, can create overheating or undercooling points; early detection of these problems can prevent major malfunctions. We are now at a stage where process engineers are using temperature measurement to help control production processes. Knowing the temporal and spatial variations of critical temperatures allows engineers to adjust setpoints and parameters of automated control systems to improve product quality and increase output. For example, plastic packaging films are produced through a continuous extrusion process. The polymer is heated above its melting point and fed into the extruder in liquid form. At this point, the polymer's viscosity is between that of sugar and water; the hotter the material, the lower its viscosity. Numerous tubes are used to feed the material, which passes through a compression press, extruding 120 to 140 feet per minute in sheets. A certain stress is maintained on the feed grid, and as the sheet passes through a cooling roller called the compression press, it is stretched to 1000 to 2000 feet per minute. This process cools, solidifies, and shapes the sheet. Of course, for a given stretching force, the viscosity of the material affects its stretch. Higher-temperature, lower-viscosity materials will stretch larger and thinner than lower-temperature, higher-viscosity materials. These sheets, exceeding 400 inches in width and only a fraction of an inch in thickness, consist of multiple zones on the feed grid, each zone having a feed tube. The final shape of the sheet is determined by the melting temperature of the feed tube. If the temperature of the feed tube is too low, the material will become too viscous and cannot be fully stretched. Ultimately, the internal stress of this material will damage the support mesh. The Competition Between Infrared Thermal Imaging and Infrared Camera Technology Although both detect the infrared thermal radiation emitted by objects at room temperature and form an image, infrared thermal imagers and infrared cameras are two very different methods. The main difference lies in the sensing element that converts electromagnetic radiation into an electrical signal output. Any object, except absolute zero, emits electromagnetic radiation. Spectral intensity (a function of frequency or wavelength) has a characteristic shape: it is 0 at 0 frequency, then rapidly reaches a peak, and asymptotically approaches 0 at infinity. As the object's temperature increases, this intensity peak shifts to higher frequencies (shorter wavelengths). For low-temperature objects (tens of Kelvin), the peak appears in the radio and microwave frequency range; for room-temperature objects (hundreds of Kelvin), the peak appears in the infrared spectral region; and for high-temperature objects (thousands of Kelvin or higher), the peak appears in the visible spectral region. [align=center]Infrared imagers use area scanning technology, simultaneously collecting infrared radiation from all parts of the image; while infrared thermal imagers use a single moving lens, moving one pixel at a time. In sequential scanning systems, the movement of the object being measured produces vertical resolution.[/align] The peak frequency of thermal radiation increases steadily with temperature, from yellow to blue, then to the ultraviolet region and beyond. Humans associate blue with cold and red with heat because of an evolutionary event: the coldest objects our ancestors encountered were water and ice, which are blue because they reflect the color of the sky; the hottest objects they encountered were wildfires, which are red or yellow. Astronomers encounter objects (celestial bodies) with temperatures ranging from thousands to millions of Kelvin, and they associate red with cold and blue with heat. The wavelengths of infrared radiation sensed by thermal imagers and infrared cameras are imperceptible to the human eye, so the instruments assign different colors to different wavelengths, creating what we call "pseudo-color images." As shown in the diagram, an infrared camera is essentially a region-scanning camera with a charge-coupled device (CCD) and an infrared filter at the front. Silicon CCD elements are highly reactive to both the infrared and visible light portions of the spectrum, so CCD cameras designed for visible light must filter out infrared light to obtain high-quality images. CCD cameras react to incident light by converting the energy of incoming light quanta into free electrons. Each incident light quantum generates a certain number of free electrons, which are collected by a circuit board – hence the name "CCD." Therefore, infrared imagers directly respond to the light quantum flux of each pixel. The camera's resolution also depends on the pixel size. Thermal imagers, on the other hand, use a device called a "thermal radiation meter" as their sensing element. The thermal radiation meter converts incoming radiation into heat, which increases the temperature of the sensing element. This temperature rise is then measured, providing a more direct reflection of the object's temperature. To generate a thermal image, the thermal imager begins by forming an image using infrared light. A static and dynamic system scans this image across the thermal radiation meter, and the output electrical signal exhibits the characteristics of raster-scanning video output. Finally, a waveform acquisition circuit (consisting of a sample-and-hold device and an analog-to-digital converter) provides the digital output. The resolution of a thermal imager depends on the dimension and sampling rate of the thermal radiation meter. The term "image sequential scanning" implies that the imager focuses only on pixels along a single line. The second dimension requires creating a two-dimensional image, which can be achieved by moving an object within the imager's field of view, perpendicular to the scan line. The essential difference between infrared thermal imagers and infrared imagers is that an infrared thermal imager plots an image as a function of position, while an infrared imager plots a density map of infrared radiation. Infrared imagers have a very fast response time, while the scan rate of an infrared thermal imager is limited by the response time of the thermal radiation meter. Therefore, the primary process control involves controlling the temperature of the molten polymer in the feed tube. In a complex system, the heater can control the melting temperature in a closed-loop manner. Because the molten polymer flows rapidly in the feed tube, there is a significant difference in material temperature between the tube wall and the internal polymer material. It is important to control the material temperature, not the feed tube temperature, so control system engineers insert probes with thermocouples into the fluid within the feed tube. Even when the control system maintains the fluid temperature at the setpoint using this method, the fluid temperature will still fluctuate from time to time. To calculate the temperature change over a certain width, each feed tube is controlled separately, so the setpoint for each feed tube (each area) on the entire substrate mesh is different. Another problem is that the molten material loses heat as it flows from the feed tube to the die, and also from the die to the compression molding machine. The temperature change from the die to the compression molding machine is a key factor causing stress. Of course, the heat loss rate is different across the entire substrate mesh, lowest in the middle and higher at the edges. Without closed-loop control, the edges of the sheet will cool and become viscous, while the center remains hot and liquid. Visualized temperature recognition by Optex Process Solutions' Andy Christie helps film manufacturers address these issues, ensuring optimized and controlled substrate mesh processes. He uses a thermal imager to monitor temperature changes in critical areas of the substrate mesh as soon as the material emerges from the die. The Raytek EC100 line-scan thermal imager has a field of view as high as a scan line, and begins vertical scanning as the substrate mesh passes through the field of view. In this way, the thermal imager scans the support mesh approximately 30 times per second along its moving direction. The speed of the support mesh divided by the scanning rate determines the position of the next parallel temperature measurement line. After the image acquisition unit acquires each individual scan line, it transmits the data directly to a laptop, which displays the scan results horizontally. Different colored dots on the lines represent different temperatures at corresponding points on the support mesh. The operator can adjust the width of the temperature range to match the temperature range of important points. Christie says, "For example, if the target temperature is 610°F, and the temperature varies from 595°F to 620°F, I would display the lower temperature of 595°F in blue, the higher temperature of 620°F in yellow, with a spectral transition in between." Temperatures above or below this range are represented in white or black. If the temperature range is 100-1,000°F, Christie expands the color to cover the entire range. As a new scan point enters the computer, the previous scan line moves up, clearing the bottom of the screen to display the new scan point. Over time, the system generates a pseudo-color image, with the horizontal axis corresponding to the direction perpendicular to the material feeder's movement and the vertical axis corresponding to the direction along the material feeder. For example, if the control cycle of a feeder heater fluctuates, the color of the corresponding vertical line will change periodically. Similarly, if a feeder continuously overheats, a vertical stripe will appear on the image, with the stripe's color corresponding to the higher temperature. This system allows process engineers to control the temperature of each feeder individually in closed loop and adjust its setpoint. For example, if the temperature at the edge of the material feeder is found to be lower than the temperature in the middle due to excessive heat loss, the engineer can increase the setpoint of the edge feeders to unify the temperature of the entire material feeder. [align=center] Figure 2: Thermal images are used by equipment engineers to monitor for early signs of accidents. The image above shows that the temperature of the motor in the middle is higher than that of the other motors. [/align] Christie says, "You can also open multiple windows at the same time; I often open another window to display the last scan." The graphical display uses a combination of quantitative data and qualitative pseudo-color images, making thermal imagers much more practical than the simple qualitative images from infrared cameras. It can provide actual measurements, while the latter can only offer descriptions like "that area is hotter," which is useful for maintenance procedures but not for control applications. Glass Tempering The process control of the support mesh is not the only example of using thermal imaging data for production process control. For example, glass tempering requires placing a whole sheet of glass at a constant temperature above its softening point and then rapidly quenching it, introducing internal stress. This method generates stress 4-5 times higher than ordinary glass, causing the glass to break into small cubes instead of large, sharp fragments when broken. Ordinary glass can be compared to a very viscous liquid, so viscous that its flow under normal forces is imperceptible. However, as the temperature rises, the viscosity of glass decreases. At the glass transition temperature (550-650 °F), with support, the glass can still maintain its shape well, but its internal atoms are already freed from the constraints of internal stress and can move freely. Clifford Matukonis, a process engineer for tempered glass quenching, explains, “After the glass sheet is heated in the furnace, it enters the air-cooling furnace.” For optimal temperature control, he transports the glass from the furnace to the air-cooling furnace via a conveyor belt at a rate of 1000-1500 inches per minute. A scanner using the same pattern as in the feedstock application monitors the glass sheet from above, with the scan lines perpendicular to the conveyor belt's transport direction, scanning parallel stripes across the entire glass sheet. Measuring More Than Just Temperature Perhaps surprisingly, infrared thermal imagers can also measure things beyond temperature. For example, the external temperature of an incomplete tank often indicates the liquid level inside. Infrared thermal imagers can be useful in certain level maintenance situations where traditional level sensors are ineffective. A thermal imager can show that the oil level in an oil-filled high-voltage insulating isolator is below a warning value; image analysis software can easily detect the problem and activate a pump to fill the low-pressure side with oil. The next image shows temperature variations caused by inconsistent humidity on a papermaking wire. High humidity accelerates evaporation. Because evaporation is an endothermic process, areas with high humidity are relatively cool. This perfectly defined effect can be quantified into humidity values ​​using temperature data from a thermal imager. [align=center] Figure 3: Thermal images can be used to measure non-thermal parameters, such as the oil level in high-voltage insulators. Figure 4: A thermal image showing the relative humidity of paper. [/align] The air-cooled furnace is a flat plate with many holes. Cold air blows upwards from the holes onto the glass sheet, rapidly "thawing" the glass and locking it in a balanced strain field determined by the annealing temperature and annealing rate. This balanced stress field contributes to the properties of tempered glass. Matukonis says, "We usually place the scanner between the furnace and the air-cooled furnace, at that point, we can find the temperature gradient across the entire glass sheet." The tempering process is controlled through a series of radiant heating elements mounted above a conveyor belt within the furnace. Each heating element has a proportional controller that uses thermocouples as sensors. The entire process control is accomplished through a PC-based system that communicates with the proportional controllers via Profibus. The heating element setpoints vary between 600-700 ℃. By changing the temperature of each heating element, the temperature consistency of the entire glass sheet can be corrected. The final temperature of this glass sheet is determined by its time in the furnace (this time is determined by the conveyor belt speed). If the conveyor belt speed is slower, the glass spends more time in the furnace, thus achieving a higher final temperature. Similar to Christie's film stacking application, thermal imaging allows Matukonis to monitor the temperature between the furnace and the air-cooled furnace, which is the most critical temperature. Currently, only thermal imaging can acquire temperature data in a spatial representation based on the temporal changes. Moreover, similar to the application of the sheet support mesh, because both the final temperature of the sheet and the temperature gradient on the sheet are known, Matukonis can adjust the heater setpoints, ensuring smooth operation of the control system. From Auxiliary to Primary Control engineers have noticed the ability of thermal imaging to optimize control loop setpoints; the next step is to remove humans from this control loop. They will develop an algorithm that enables process control to operate automatically, while increasing the number of control levels to make automated production systems more robust. Thermal imagers are used in factories as a primary sensor for obtaining quantitative information, helping engineers adjust process control. Future systems will equip thermal imagers with programmable automatic controllers (PACs) to calculate appropriate setpoints and transmit them to the PID controllers of each process control unit. These thermal imagers collect temperature data in both the spatial and temporal domains, while the PACs analyze this data and transform it into more direct process parameters, such as slope, liquid level, and control system stability. This system will optimize the entire production process, maximizing both quality and output.