Objectives  The inherent capacity of ship system is an important part of efficiency. In order to evaluate and research inherent capacity of ship system,

  Methods  for the function indexes which are difficult to be quantized, the quantitative correspondence between fault status and the function indexes is established by the external test point test and reliability model on the basis of the index system of ship system. A multi-classed robust energy model least squares twin support vector machine (RELS-TSVM) method is used to make a preliminary evaluation for the system inherent capacity.

  Results  The simulation result show that the multi-classification evaluation model proposed in this paper improves the classification accuracy by 8.97% compared with the model of simple support vector machine. With the help of added system-level test plan, the coverage ratio of system faults is improved, the credibility of evaluation results is increased, and the problem of data acquisition difficulty in data-driven inherent capacity evaluation methods is solved.

  Conclusions  The feasibility and advantage of ship system inherent capacity evaluation scheme based on test analysis and RELS-TSVM are verified. This paper provides a new idea for weapon equipment efficiency evaluation and research.