Data Driven and Fuzzy Synthetic Evaluation Based Submarine Cable Health State Assessment

Abstract

—Health state assessment is of great importance to ensure the safe and stable operation of submarine cables. Submarine cables have composite structural characteristics and work in complex and influential environments, which hinders manual inspection, arouses difficulties in maintenance, and causes challenges on health state assessment. To this end, a data driven and fuzzy synthetic evaluation based submarine cable health state assessment method is proposed in this paper. Firstly, following the technical guidelines of submarine cable health state assessment, dynamic submarine cable online monitoring data and static data including periodic preventive test and operation inspection results are integrated to build a three-layer index system. In addition, long and short-term memory (LSTM) neural network based ultra-short-term prediction models are built to obtain the predictions of online monitoring indices. On this basis, together with the static data, a fuzzy synthetic evaluation method is used to establish a two-stage state evaluation process to obtain the multi-period health states of submarine cables. The effectiveness of the proposed method is verified with an actual submarine cable from an offshore platform in an oilfield of China National Offshore Oil Corporation (CNOOC).