Text mining and association rules-based analysis of 245 cement production accidents in a cement manufacturing plant

Abstract

Abstract Accidents such as collapse, fire, suffocation, poisoning, scalding and mechanical injuries occur frequently in cement industries. Understanding the causes of past accidents in cement companies is essential to prevent cement production accidents and reduce safety risks. However, there is currently no cause analysis of accidents that have occurred in cement companies. Hence, this paper takes cement accident cases as the basis of research, proposes a unified report analysis framework, combines data mining technology, probes deeply into the law of cement production accidents, and establishes cement accident causation analysis model to provide a basis for current safety management decisions. Firstly, 245 accident records were collected to categorize the causal factors of cement accidents in this plant according to the LDA model, and then a systematic accident causal analysis method was proposed according to the 24Model to establish a unified report analysis framework. Based on this, an improved Apriori algorithm suitable for multi-dimensional multi-layer cement enterprise accident correlation rule mining was proposed to improve the efficiency of accident mining. Using the improved Apriori algorithm, the correlation between accident causative factors and accident types as well as accident causative factors of cement enterprises was quantitatively mined, and targeted safety management suggestions were put forward.