Abstract
<title>Abstract</title> <p>The security state awareness is crucial for petrochemical industrial control network equipment, as it directly affects the operational efficiency and security performance of petrochemical industrial control systems. However, frequent data exchange in these network makes petrochemical industrial control network equipment inevitably face concerning security issues. In response to the high complexity of situational awareness and delayed transmission of security information in the petrochemical industrial control network, we propose a deep learning-based security state awareness scheme for petrochemical industrial control network equipment. For dynamic and continuous situational spaces, this scheme constructs a security state awareness model for petrochemical industrial control network equipment. We design a hybrid awareness model combining temporal convolutional network and multi-layer gate recursive units (TCN-MGRU) to detect the security state of petrochemical industrial control network equipment. The model features a low complexity classification process, which markedly improves awareness accuracy. Finally, the experimental results verify that TCN-MGRU significantly enhances the accuracy of security state awareness for petrochemical industrial control network equipment and also confirm its applicability. For petrochemical industrial control network equipment, it provides robust support for the security protection.</p>