According to the EU study on electricity supply disruptions, in the period 2010–2014 up to 850 GWh of electricity annually have not been supplied to the consumers. The lost value from these disruptions is estimated to be up to 25 billion EUR per year to the commercial users. The Lithuania annual average of the power lost due to disruptions is 410 minutes for single user which significantly exceeds the average of all member states (136 min.). The vast majority of these disruptions occur due to the problems in distribution grid at the distribution or transmission level. Lithuania counts more than 200 high voltage transformers (110 kV), which are distributed across the country. It’s estimated that the replace of malfunctioned transformer can cost around 300 thousand EUR, while if the cause of malfunction is detected at early stage, such expenses can be reduced 10 times. According to the statistics, around 40% of unplanned disruptions are caused by failure of equipment or material damage. The material damage is explained as the voids and cracks in the insulation layer of the conductor which leads to the partial discharge phenomenon (PD). The PD is defined as localized breakdown of insulation layer under the stress of high voltage. These voids in the insulation layer tend to develop over time, which after a while leads to the full discharges and false operation of the equipment. The detection and localisation of the partial discharges at early stage is of vital importance to ensure safe and reliable supply of the electric power. In this project, the partial discharge detection and localisation method is being proposed which uses the fusion of ultrasonic, optical and electromagnetic field measurement methods along with the signal processing algorithms based on artificial intelligence. At the end of the project, the multichannel ultrasonic system augmented with the special signal and image processing methods will be developed and tested for partial discharge detection in different environments.
Project funding:
KTU R&D&I Fund
Project results:
The prototype partial discharge measurement system was developed throughout the project that is designed to operate in air and to assess dielectric conditions of connectors of air power lines and bushing insulators. In contrast to other commercially available systems, the proposed prototype has the following advantages:
• uses multiple ultrasonic channels for discharge detection and localisation;
• does not require manual steering of the apparatus to find the source;
• requires no physical connection to the HV assets, making an easy, non-invasive and cost-effective inspection without disturbance to the power network;
• can be used for permanent or temporal monitoring of discharge activity with on-site deployment. Clustering of multiple systems in a client server architecture is possible as well.
• incorporates sophisticated ultrasonic signal processing algorithms, machine learning and deep learning methods for autonomous discharge detection with increased reliability and accuracy;
• does not require additional software or transformer models to obtain discharge source position.
Period of project implementation: 2019-04-01 - 2019-12-31
