(7b) Improving Furnace Operation Using Visible-Spectrum Embedded Cameras

If one were to ask anyone who has looked into an ethylene furnace during operation, they would likely agree that there are many things related to the operation of the furnace that can be qualitatively observed by a person. This includes identifying sections of tubes that are hotter than other sections, whether or not burners are operating correctly, if tubes have shifted or broken, and amongst other items. All of this can be observed simply by the human eye when properly protected. While this information is useful, it is normally only qualitative and something a person can only observe whenever they open a small door on the furnace to look inside, which can also cause other issues.

This paper explores the application of visual-spectrum embedded furnace cameras and advanced image processing technology to quantitatively measure this information and more, in real time, during furnace operation.

Embedded furnace cameras are designed to be permanently installed, without a retraction system, into the refractory of a furnace. This significantly reduces the required amount of cooling utilities for the camera and minimizes problems incurred even if cooling utilities are lost for a period of time.

Traditional tube metal temperature monitoring is done using thermocouples or using a point IR measurement device. There are many issues with these methods, including the expense and failure of thermocouples, and the repeatability and safety of operators opening doors to make measurements with a point IR device. A better method for this measurement would be to use a camera. Many cameras or sensors that are installed to monitor the inside of furnaces use infrared sensors intended to measure the temperature of a small area. While these sensors can provide helpful information, they can also be subject to various issues including accuracy problems stemming from emissivity. Within the operating temperature ranges of ethylene furnaces, the solid materials in furnaces will emit visible spectrum light due to blackbody radiation. This visible spectrum light can be measured to determine temperature similarly to how an IR sensor measures light in the IR spectrum. In the visible spectrum, however, emissivity concerns become negligible. Using higher resolution sensors with lens optics that cover a wider field of view, it is also possible to see large sections of many tubes all in one image. This allows the measurement of average, minimum, and maximum temperatures on multiple tubes and the identification of hotspots, all at the same time with a single camera.

While a person looking into a furnace can qualitatively say that a tube has shifted over time, an embedded camera turns that visual into a repeatable measurement. Advances in imaging analysis allow for tubes to be intelligently identified and tracked by software. This information can all be fed into control systems to alarm to problems with tubes, such as if tubes move outside of an allowable range or if a tube breaks.

While all of these different measurements are being made, a visible spectrum camera also provides the image which would mimic what the human eye sees inside of the furnace. This image provides a constant visual verification of all measurements in a safe and reliable way that can limit or eliminate the need for a human being to open a door and expose themselves to the process inside of the furnace. Additionally, the video can be recorded and saved to go back to if there is ever a question in the future about an event that occurred during furnace operation.

By utilizing embedded visible-spectrum cameras and image analysis, ethylene producers can reliably measure more than ever before with the confidence that they have the images to back up their data.