Application of NARX neural network for predicting marine engine performance parameters

Though the maritime industry is still predominantly reliant on a time-based, prescriptive approach to maintenance, the increasing complexity of shipboard systems, heightened expectation and competitive requirements as to ship availability and efficiency and the influence of the data revolution on vessel operations, favour a properly structured Condition Based Maintenance (CBM) regime. In this respect, Artificial Neural Networks (ANNs) can be applied for predictive maintenance strategies assisting decision makers to select appropriate maintenance actions for critical ship machinery. This paper develops a Nonlinear Autoregressive with Exogenous Input (NARX) ANN for forecasting future values of the exhaust gas outlet temperature of a marine main engine cylinder. A detailed sensitivity analysis is conducted to examine the performance and robustness of the NARX model for variations in the time series data, demonstrating virtuous performance and generalisation capabilities for forecasting and the ability to employ the model for monitoring and prognostic applications.