Health status assessment for ship diesel engines based on LSTM prediction and cloud barycenter model

Objectives In response to the development needs of smart engine rooms on ships, this paper proposes an assessment method for the health status of ship diesel engines. The method is based on long short-term memory (LSTM) neural network prediction and cloud barycenter evaluation, aiming to enhance the operational and maintenance capabilities of the engines.

Methods First, an evaluation indicator parameter set is constructed based on the deviation between LSTM-predicted and measured parameters. Then, the analytic hierarchy process is used to construct parameter weights, and the cloud barycenter evaluation method is employed to assess the health status of the diesel engine. Finally, tests are conducted using actual ship diesel engine data from both the early normal and later degradation periods.

Results The results indicate that the evaluation value of the diesel engine in the early normal state is 99.94 (healthy), while in the later degradation state, it is 81.71 (good), achieving the goal of health status assessment.

Conclusions The proposed method can be applied to the health status assessment of ship diesel engines and other power equipment, offering practical application value.