Prognosis of Connector Disconnection Using a Canary-Based Approach

The electrical connector disconnection is a common problem in automotive systems. It can be caused by bad design, manufacturing issues, ageing in harsh environment, or low product quality. An imminent disconnection during driving may result in severe safety issues. A loose connection manifests itself as intermittent faults of various vehicle components, which is hard to diagnose and likely leads to unnecessary component replacement or dealership revisit. In order to predict the connector disconnection, a low-cost canary-based approach is proposed in this paper. A shortened male terminal is employed to foretell the loose factory terminals in the same connector housing. The dimension and placement of the shortened terminal are theoretically and experimentally investigated to achieve optimal performance. The proposed solution is tested and compared to other common diagnostic and prognostic approaches, including inductance–, capacitance–, resistance–based approaches, time domain reflectometry, and frequency domain transmissometry, using a connector bench test setup. The placement variation test and the accelerated vibration test are performed to simulate the long-term real driving scenario as well. It’s shown that the proposed solution is capable ofpredicting connector disconnection robustly before the vehicle functionality is affected.