Research and Development of a Corrosion Monitoring System for Marine Equipment

Abstract A corrosion monitoring system utilizing a resistive probe for marine engineering equipment has been designed, incorporating various circuit structures, including biofouling prevention, temperature compensation, corrosion rate measurement, and data processing and transmission circuits. On the software side, a microcontroller-based lower-level program and a PC-based upper-level software were developed to facilitate the transmission and display of corrosion data. To overcome the limitations of traditional ARIMA time series forecasting models, a novel BP-ARIMA corrosion prediction model, which integrates BP neural networks was proposed. The performance of this BP-ARIMA model was evaluated using corrosion rate data from salt spray tests on 316 stainless steel specimens. The results demonstrated an enhancement in prediction accuracy compared to both the ARIMA model and standalone BP neural network models.