LLK sought to act within the M-START to reduce the impacts caused by tears in the conveyor belt. Current tear detectors are not effective, as they need a drop of material in order to function. The great occurrence of false alarms and the high response time of this type of system generates enormous losses for the mining company due to the loss of carpet and the stoppage of production.

Site: LLK
Cycle: M-Start Cicle2
Mining Company: Vale

Solution

As a solution to the challenge presented, LLK proposed a new system for the detection of tears in belts with low response time and without the occurrence of false alarms, called RADEC Visão. Using computer vision, our solution analyzes the relative distance between the edges of the belt to detect the occurrence of tears even without falling material. During the PoC period, we aimed to develop and test the system on a full-scale prototype conveyor, developed by LLK exclusively for the M-START. Tests were carried out with and without load, and the main types of possible tears were simulated: tear with opening, tear with overlap, edge tear, in addition to belt misalignment.

Results and advantages

The system showed excellent results in the tests during the proof of concept, proving its good functioning. The response time obtained was less than 1 second and the system did not present any occurrence of false alarm in all tests performed. In addition, the system efficiently detected edge tears, with an opening and overlap (without material drop) of less than 15 millimeters, which represents 1% of the width of a 48-inch belt. Another point observed in the tests is that the RADEC Visão is also capable of identifying belt shifts to the left and right, confirming its functionality for both tear detection and misalignment.

Team

Our team is formed by engineers and engineering interns from the areas of control and automation, systems and mechanics. With experience in innovation, we have specialists in computer vision systems, using different types of cameras, lenses, lighting and software in our solutions, as well as monitoring 4.0 of equipment with different types of instruments, such as piezometers, inclinometers, accelerometers, among others .