论文部分内容阅读
这种系统从各种机器的若干部位采集有关状态的数据,并运用神经网络学习表征正常工况的征兆。所有采集到的新数据均与知识库中的征兆作比较。如果新的数据是系统已知的,可立即将诊断结果返回用户。当出现新的情况时,机器监测专家库即处于待命状态,运用电子邮件和GSM短信息服务作出诊断,然后将其返回与原始数据联接。经过再训练之后,系统将能识别所有类似的数据。
This system collects data about the state from several parts of various machines and uses neural networks to learn signs that characterize normal conditions. All new data collected is compared with the symptoms in the knowledge base. If the new data is known to the system, the result of the diagnosis can be immediately returned to the user. When a new situation arises, the machine monitoring expert bank is on standby, making a diagnosis using e-mail and GSM short message service, then returning it to the raw data connection. After retraining, the system will recognize all similar data.