Abstract: Internet of Manufacturing Things obtains real-time state data streams through the deployment of massive sensor nodes in industrial processes. Using the complex event processing method to analyse and process the large-scale and real-time data streams generated in the monitoring of production process monitoring is necessary, and to judge the stability of the process flow is one of the important problems concerned with production department. However, the raw data from the physical world has high level of uncertainty because of the inherent inaccuracy of sensor readings, bringing greater difficulties to judge the stability of the process flow. The probability of the event model is defined and the probability event detection method used based on finite automata with buffer uncertainty (NFA with run buffer, rNFA), and a method combined with the runs test and the grey system model and the bootstrap experience prediction is put forward to make quantitative checking of the stability of the process flow. Experimental results show that the proposed method can effectively detect probability event flow state of the global production.

Key words: process flow, complex event detection, runs test, the grey system model, bootstrap prediction