Communication problems are directly or indirectly related to the causes of industrial accidents. This may be associated with problems in information management at the operational level. Before evaluating information management on the shop floor, a documentary analysis was carried out in this study to identify problems in the communication among the causes of 5 accidents in the oil sector in Brazil. After documentary analysis, an adaptive neuro-fuzzy inference system (ANFIS) was developed to assess information management on the shop floor. A case study was developed with a sample of 120 respondents through the application of a survey. Managers, supervisors, and operators from a land-based oil production region participated in this study. The documentary analysis identified communication problems in 53% of the reports. The neuro-fuzzy model performed well, with the root mean square error (RMSE) being 0.229 in training and 0.296 in data verification. The results suggest that the ANFIS method can be successfully applied to establish a model of analysis of information management on the shop floor.

中文翻译：

---

用于评估车间信息管理的神经模糊系统的开发

通信问题与工业事故的原因直接或间接相关。这可能与操作层面的信息管理问题有关。在评估车间信息管理之前，本研究进行了文件分析，以找出巴西石油行业 5 起事故原因之间的沟通问题。经过文件分析，开发了自适应神经模糊推理系统 (ANFIS) 来评估车间的信息管理。通过应用调查，对 120 名受访者进行了案例研究。来自陆地石油生产区的经理、主管和操作员参与了这项研究。文件分析在 53% 的报告中发现了沟通问题。神经模糊模型表现良好，均方根误差 (RMSE) 在训练中为 0.229，在数据验证中为 0.296。结果表明，ANFIS方法可以成功地应用于建立车间信息管理分析模型。