High Voltage DC Arc Detection Model

• Main Author: Farag, Bassem
• Format: Others
• Language: English
• Published: Uppsala universitet, Institutionen för informationsteknologi 2018
• Subjects: Engineering and Technology; Teknik och teknologier
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-358478

High voltage (HV) battery systems are widely used in many applications nowadays. And due to the safety concerns regarding lithium-ion cells, the safety of the lithium-ion based battery systems is vital. One crucial danger for lithium-ion cells is heat, And as arcs formation can lead to heat generation within the system, it is important to detect arcs that take place frequently within HV battery systems. This thesis is done in cooperation with Northvolt AB and it focuses on assessing the ability to detect the occurrence of arcs in the system, but it does not focus on preventing arcs.The goal is to build a detection system to identify the occurrence of arcs, both within the battery system and in the connection between the battery system and the load (vehicle). The detection circuit should not affect the ability of the isolation measurement unit inside the system, and the detection system should be protected at all times. The circuit was designed and tested using LTSPice software. This is due to the absence of a ready system to test the circuit against at Northvolt. The system was able to detect arcs both within the battery system and when connecting the battery system to the vehicle. Additionally, as required by Northvolt, the detection system is designed without affecting the isolation measurement unit and the detection system is kept safe at all times by using an isolation circuit. Future work is recommended to generalize the detection system so it can be used in different high voltage applications. This can be done by testing the system against other HV systems and updating the filter and amplifier’s values, as well as the software thresholds. Additionally, it is recommended that the software module is calibrated against the real system during hardware testing. This calibration will optimize the software module and, thus, result in better detection.