Adaptive robustness evaluation for complex system prognostics and health management software platform

With the quantity and complexity of various complex systems have increased, their safety and reliability face many challenges. The prognostics and health management (PHM) software platform for complex systems is one of the important tools to ensure the reliability and safety of them. On the other hand, due to the uncertainty of sensing data, the robustness of PHM software platforms needs to be evaluated to ensure their performance. However, it is highly subjective and time-consuming for the existing robustness assessment methods which rely on expert experience. To solve these problems, a novel adaptive robustness evaluation method with Genetic Algorithm (GA) and Random Forest Algorithm (RFA) for the complex system PHM software platform is proposed. Firstly, the basic robustness indicators are extracted based on the classical metrics of PHM software platforms. Secondly, more sensitive indicators to the robustness of PHM software platforms are selected by GA from basic robustness indicators. Then, the classification is conducted by the selected indicators through RFA. Finally, the robustness of PHM software platforms is evaluated by the classification accuracy. Experiments with simulation data show that the proposed method has better performance, which is suitable for the robustness evaluation on the PHM software platform of complex systems.