We have built a unique computer vision-based tool for cement kiln operations, deployed at multiple client sites in India.

The tool takes the live input camera feed from the cement kiln and basis on the computer vision algorithm running on the camera feed, identifies whether the cement kiln is healthy, dusty, or hot. As this is a continuous process without human intervention, it leads to a reduction of the time when the kiln is dusty or hot and improves the share of time when the kiln is healthy.

In addition, the tool also stores the kiln images (at 2 min intervals) on the cloud and they can be retrieved easily via our web application. This facilitates root cause analysis in areas such as refractory performance amongst others.

In addition, the tool also captures the data of different process parameters and thereby recommends the right fuel rate via neural network models.

Impact:

• Better visibility into kiln conditions to prevent jamming

• 3% reduction in fuel rate through better control