Human error assessment in start-up operations of utility systems in a small chemical plant using CREAM

Start-up operations in small chemical plants represent a critical yet underexplored phase for human error analysis. This study presents the first systematic application of the Cognitive Reliability and Error Analysis Method (CREAM) to assess human reliability during start-up operations of three utility systems steam boilers, water-cooling pumps, and air compressors in a small Malaysian chemical plant. Unlike existing studies that focus on routine operations or large-scale facilities, this research addresses the unique challenges of manual start-up procedures in resource-constrained environments. Both basic and extended CREAM versions were applied using Hierarchical Task Analysis (HTA) validated by seven experts with more than 10 years of experience. The analysis revealed that all systems predominantly operated under tactical control mode, with human error probabilities ranging from 0.073 to 0.121. Water-cooling pump operations showed the highest risk (0.320) due to time constraints and collaboration quality issues, while boiler operations demonstrated the lowest risk (0.014) through structured procedures. Critical failure modes were identified in observation and timing-related tasks, particularly in speed verification and parameter adjustment subtasks. This study demonstrates CREAM's applicability to small-scale chemical plant start-ups and provides quantitative reference values for integrating human reliability assessment into Process Safety Management (PSM) systems. The findings support targeted interventions including procedural standardization, enhanced training for high-risk subtasks, and improved shift handover protocols to reduce human error in early-phase operations.