A complete bearing fault diagnosis process includes the following five aspects:
(1) signal measurement. According to the working environment and nature of the bearing, select and measure signals that can reflect the working condition or state of the bearing.
(2) Feature extraction. Through certain signal analysis and processing methods, useful information that can reflect the bearing state is extracted from the measured signal.
(3) National recognition. Identify the bearing state by some state identification method, that is, simply judge whether the bearing is faulty.
(4) State analysis. According to the symptoms, further analyze the situation and development trend of the state. When a fault occurs, analyze the fault type, nature, location, cause and trend in detail.
(5) decision-making intervention. According to the bearing state and its development trend, make decisions such as adjustment, control or continuous monitoring.
The purpose of bearing fault diagnosis is to determine the degree of fault from the fault location to the fault nature. Because neural network has the ability to deal with complex multi-modes and the functions of association, speculation and memory, it is suitable for fault diagnosis of ball bearings.
Using neural network to diagnose the fault of rolling bearing can send out early warning signal when the fault occurs, repair or replace the bearing that will fail in advance, shorten the downtime and reduce the maintenance cost, so as to minimize the loss and ensure the stable and safe production.
……
Please send an email to young _ tower @ 126.com for detailed answers.