Real-time knowledge discovery for process objects

This paper designs a real-time knowledge discovery method which is used for production data of thermoelectric steam boiler. For the growing stream of industrial data conduct knowledge discovery in real time, it can obtain the latest data trends of each link constantly. The recent steam boiler data is predicted and simulated by model prediction. This method can get the latest implicit knowledge from the updated data. In order to better assist the adjustment of equipment in thermoelectric production, the method mainly includes two parts, one is the establishment of the initial model, and the other is the incremental model update. The first part includes data preprocessing, link clustering, association rule chain mining, modeliig and prediction. The second part includes new data stream preprocessilig, data stream clustering, association rule chain mining and model updating. Through the algorithm design, Knowledge can be acquired more effectively. In practical applications, it becomes more timely detection of production faults, better decision-making in production. It helps thermal power plant energy saving, emission reduction and enhance production safety.