In the United States, approximately half of the electricity is generated from coal-fired boilers. Efficient power production requires plant operators to monitor boiler tubes and slag deposits that prevent boilers from running at peak efficiency. Proactively maintaining boilers helps coal plant operators to improve boiler uptime by reducing the number of manual cleaning cycles and decreasing tube erosion that may be caused by excessive cleaning.
"In the past," says Lenny Shaver, director of business development with LumaSense Technologies (Santa Clara, CA, USA), "this process was performed manually by operators who viewed the combustion process and internal components of the boiler through observation ports. Not only is this not accurate or repeatable, it is also hazardous since boiler temperatures are often as high as 2000°F." What makes it even more complicated is that the human eye is sensitive to light from between 400 and 700 μm, so boiler tubes cannot be accurately visualized during the combustion process.
To overcome these problems, LumaSense has developed a turnkey system known as BoilerSpection for continuous infrared (IR) imaging inside the boilers (see figure). The system uses a vanadium oxide-based, 320 × 240-pixel microbolometer camera with 0.08-K sensitivity, equipped with a 3.9-μm bandpass filter, to image the boiler tubes during the combustion process.
Of course, manufacturing a camera to operate in such harsh environments is not a trivial task. The imaging system is fitted with a 16-in. borescope that is placed through the observation port of the furnace. The camera system is then mounted to this external port. Cooled air is passed through the borescope and the camera housing to lower the temperature of both the lens and camera to 2000°. This also allows the lens and camera system to be purged of any contaminants that may adversely affect the system's operation.
Images captured at up to 30 frames/s are transferred either digitally over an Ethernet interface to a host PC or via an NTSC interface to a color monitor. In this way, images can be processed and displayed on the host PC and simultaneously viewed in real time. After images are transferred to the host PC, they are processed to determine specific characteristics.
To check the buildup of ash and slag, for example, specific regions of interest (ROIs) within each image can be programmed so the operator can monitor how the temperature within them changes over time. Based on these comparisons, the system can be used to automatically alert an operator, trigger an alarm, or control cleaning equipment on the boiler.
By monitoring these images over a specific time period, the system also provides historical data of how the coal burning process is performing, which helps to maximize the efficiency of the process. Historical data are also crucial for evaluating the effect of other changes to the process, such as new fuel sources or maintenance to the burners or coal pulverizers.
To further extend access to images captured by the IR cameras, the host PC software broadcasts the images as streaming JPEGs to a web page. Any user that is granted access on the plan network can view the images using a web browser. Access is controlled by standard Internet Information Services (IIS) tools that the plant's IT department controls.
The system is already being used by a number of coal-powered facilities. One such facility, Alabama Power's Miller Steam Plant, has deployed numerous cameras to monitor two of its four 720-MW boilers. The 12-stories-tall, 60-ft-wide boilers each use six BoilerSpection cameras to monitor the combustion process at various locations.
Prior to using BoilerSpection, the Alabama plant's operators estimate they performed approximately 51 de-slags on the boilers in 2010. During a six-month period using the BoilerSpection system, only 12 of these processes were performed, resulting in huge savings to the site. The plant is now planning to outfit its remaining two boilers with the vision system.