Machine learning-based industrial equipment fault preventive recognition method

Abstract

The invention discloses a machine learning-based industrial equipment fault preventive recognition method. By the adoption of an artificial intelligence algorithm of machine learning, a state forecasting model is continuously learned and updated, so that a working state of industrial equipment can be monitored and forecast in real time, and a user can find an abnormality of the industrial equipment at the first time and sound an alarm in time to avoid an accident. According to the machine learning-based industrial equipment fault preventive recognition method, a typical fault case in the whole full life circle of the industrial equipment is used as a learning object, and furthermore, by the combination of file data of the industrial equipment, and reliability data of a regional environment condition, particularly a key element, are used as the learning object to estimate the reliability of the industrial equipment from multiple angles, so that a forecast result is more accurate.

Description

technical field [0001] The invention relates to a machine learning-based preventive identification method for industrial equipment failures. Background technique [0002] The development of industrial equipment has gone through the stage of digitalization and informationization, and is now moving towards the stage of intelligence. The emergence of intelligent industrial equipment makes it possible to realize preventive maintenance technology. [0003] Smart manufacturing based on information systems such as smart equipment and smart factories is leading the transformation of manufacturing methods: collaborative design, precise supply chain management, and full life cycle management are reshaping the industrial value chain system. In the stage of the fourth industrial revolution, the full life cycle management of industrial equipment is becoming more and more important to users. [0004] Artificial intelligence originated around 1950. After decades of development, with the ...

AI Technical Summary

Problems solved by technology

Whether it is the equipment's built-in detection software or manual regular maintenance or a combination of the two, it mainly depends on the knowledge and experience of the maintenance personnel, which requires high skills for the maintenance personnel, and it is still impossible to accurately predict the point and location of the equipment failure. time

Industrial equipment is often sold all over the country and even abroad. It takes a lot of manpower and financial resources to rely on technicians to regularly inspect and maintain

In the absence of effective measures for preventive maintenance, sudden failure of industrial equipment will cause losses to production and operations. If the cause of the failure cannot be diagnosed or key components are out of stock, the loss will be even greater if it cannot be repaired as soon as possible.

[0009] The disadvantages of the existing technology are: 1. It is limited to the electronic informatization of the existing workflow and data, and the work efficiency has been improved to a certain extent, but the component of manual participation is still very large; 2. It is limited to the maintenance of industrial equipment within the enterprise. The amount is too small to realize artificial intelligence; 3. It is limited to the data of the device itself, without considering the influencing factors outside the device, and the deviation of the prediction result will be large

There are also some studies on fault prediction and identification, but none of them are well combined with the application of preventive maintenance of industrial equipment, and the practical value is not great

Images