Alliander is a company that manages power grid connections to about 3.3 million clients. Switching from fossil fuels to renewable energy puts an additional load on the power grid. Therefore, Alliander must expand the capacity of the power grid. Expanding the power grid entails the construction of power connections over long distances, which are realized by splicing voltage cables. Splicing medium voltage cables is labour-intensive, so more qualified technicians are needed to improve and expand the power grid. Unfortunately, technicians are hard to come by due to the tension in the job market.
Before splicing two medium-voltage cables, a technician must prepare the cable ends. Part of this preparation process is peeling the insulation shield from the insulator. The Cable Splicing Robot project at the University of Twente aims to automate part of the splicing process. This cable splicing robot project involves a control system. This control system needs a monitoring system to measure the peeling process. This thesis project focuses on developing this monitoring system.
Peeling the insulator shield off the cable results in a chip. This chip resembles information about the peeling process. So, the monitoring system observes the chip and extracts information about the peeling process using computer vision. This information is suitable as a measured value for the cable splicing robot.
Validation results show that the monitoring system deviates less than 4 \% from readings manually measured with a calliper. Technicians monitor the chip by visual inspection, which is much less accurate than using a calliper for measuring the chip. So, the monitoring system proves to be valid concerning accuracy. The monitoring system must also deliver measurements fast enough for the control system to function well. The sample time is 105 milliseconds. The monitoring system also proves valid concerning sample time.