Workplace Safety Assessment Using Adaptive Neurofuzzy Inference System Approach with Partial Least Squares Structural Equation Modeling - Case Study of Sarkhun and Qeshm Gas Refining Company

Document Type : Original Article

Authors
Saeed Salimi 1
Shahla Paslar 2
Moslem Alimohammadi 3
1 Islamic Azad University, Bandar Abbas Branch, Faculty of Engineering and Technology, Department of Industrial Engineering
2 Assistant Professor, Department of Industries, Faculty of Technology and Engineering, Islamic Azad University, Bandar Abbas, Iran
3 Industrial Consultant, Head of Safety Unit, Sarkhun Gas Refining Company, Bandar Abbas, Iran
Abstract
This research utilized an Adaptive Neuro-Fuzzy Inference System (ANFIS) integrated with Partial Least Squares (PLS) structural modeling to evaluate perceived workplace safety, defined as employees' willingness to adhere to safety regulations. A survey methodology was employed, measuring six key components: tacit safety knowledge, explicit safety knowledge, psychological safety attitude, emotional safety attitude, behavioral safety culture, and psychological safety culture. The study was conducted with 90 employees from Sarkhun and Qeshm Gas Refining Company. Data analysis was performed using SPSS, Smart-PLS, and MATLAB software. Findings revealed that all six identified factors significantly impact perceived safety. Additionally, the ANFIS modeling results specifically demonstrated that behavioral safety culture is the most crucial predictor of an employee's perceived safety in the workplace, highlighting its paramount importance. This research utilized an Adaptive Neuro-Fuzzy Inference System (ANFIS) integrated with Partial Least Squares (PLS) structural modeling to evaluate perceived workplace safety, defined as employees' willingness to adhere
Keywords
Workplace safety
safety culture
structural equation method
partial least squares
fuzzy inference system

  • Received Date 18 September 2025
  • Received Date 08 December 2025
  • Accepted Date 07 February 2026
  • Published Date 21 January 2026