Dynamic detection of major equipment defects in petrochemical enterprises plays an important role in maintaining energy production safety and stability. The current traditional detection methods lack accuracy and real-time performance, making it difficult to effectively meet the demand for dynamic defect detection in large-scale energy production. In order to improve detection accuracy and efficiency, achieve safe and reliable energy production and construction, this article combined big data analysis technology to conduct in-depth research on risk assessment and optimization of dynamic defect detection of major equipment in petrochemical enterprises under operating conditions. This article first analyzed the influencing factors of major equipment defects in petrochemical enterprises, and then conducted in-depth research on risk assessment and optimization of dynamic defect detection with the goal of ensuring equipment operation safety and stability, combined with big data analysis technology. Finally, the risk of dynamic defect detection in petrochemical equipment during operation was evaluated and optimized through clustering algorithms. To verify the effectiveness of the method proposed in this paper, in the experimental analysis, the traditional Convolutional Neural Network (CNN) was used as a reference to conduct experimental analysis from two aspects: risk identification accuracy and evaluation efficiency. The results showed that in the efficiency analysis of risk assessment, compared to the CNN method, the average completion rate of risk assessment using the big data analysis method in this article increased by 9.01%. The conclusion indicates that big data analysis methods can quickly and accurately identify and evaluate the risks of major equipment in petrochemical enterprises, which helps to promote the development of energy production safety.