Real time measurement of slag volume in a pellets plant

685 million tons of pellets are produced annually, 33% of which are produced with pellet plants of the grate-kiln type. The pellet plants of the grate-kiln type are advantageously operated by burning solid fuels, usually coal. The ash from the fuel together with the dust from the pellets creates a sticky and problematic slag that sticks to the walls inside the oven. When the slag coating becomes too thick, it unpredictably detaches and damages the machine or leads to an involuntary maintenance stop to remove it.

In previous projects, the development of camera systems to monitor the inside of the furnace has proven to be a successful strategy for detecting the most obvious problems inside the furnace. Using the images from these cameras, image analysis has also enabled quantitative analysis of the slag growth. The disadvantage of the camera surveillance that is used today is that it is difficult to perceive how large the build-up is.

The aim of this project is to develop an improved and quantitative method using stereo camera technology and AI methods for determining the volume of slag build-up. With the help of this system, maintenance downtime should be reduced, by being able to predict when the slag formation becomes too large and needs to be removed. The system must also be able to help the operators to understand which operating conditions lead to rapid build-up of slag and how the build-up of slag can be minimized. Through less unplanned stops and new knowledge about the process, it can be assumed that the energy consumption and related emissions per pellet will decrease.