Power device with self-health status prediction function and self-health status prediction method thereof and cloud server suitable for a plurality of power devices

A cloud server suitable for a plurality of power devices, the cloud server comprising:

• a communication module;
• a database, configured for storing basic information, sensing data, health status prediction data, abnormal event data, lifetime data and maintenance data of each power device, and for storing a health prediction index threshold corresponding to a model name of each power device;
• a prediction-model training module, configured for constructing and training a health index prediction model for each model name, wherein each health index prediction model is constructed and trained, by way of machine learning, based on the basic information, the sensing data, the health status prediction data, the abnormal event data, the lifetime data and the maintenance data of the power devices with a corresponding model name in the database;
• and
• a data collection module, configured for collecting basic information, sensing data, health status prediction data, and abnormal event data of each power device through the communication module, obtaining lifetime data of each power device based on corresponding basic information, providing a web-based user interface for collecting maintenance data of each power device and the health prediction index thresholds, and storing all of the collected data, the health prediction index thresholds, and all of the obtained lifetime data to the database,
• wherein the steps performed by the prediction-model training module in constructing each health index prediction model comprise:
• obtaining the basic information, the sensing data, the health status prediction data, the abnormal event data, the lifetime data and the maintenance data of the power devices with a corresponding model name from the database, and constructing a failure probability prediction model by way of machine learning accordingly;
• putting the obtained basic information, sensing data, health status prediction data, abnormal event data, lifetime data and maintenance data of each power device into the failure probability prediction model to obtain a predicted failure probability of each power device;
• constructing a remaining lifetime prediction model by way of machine learning based on the obtained basic information, sensing data, health status prediction data, abnormal event data, lifetime data, maintenance data and predicted failure probabilities of the power devices;
• putting the obtained basic information, sensing data, health status prediction data, abnormal event data, lifetime data, maintenance data and predicted failure probability of each power device into the remaining lifetime prediction model to obtain a predicted remaining lifetime of each power device;
• and
• constructing a health index prediction model by way of machine learning based on the obtained basic information, sensing data, health status prediction data, abnormal event data, lifetime data, maintenance data, predicted failure probabilities and predicted remaining lifetimes of the power devices.