Equipment health assessment and fault-early warning algorithm based on improved SVDD

With the rapid development of Internet-of-Things and big data, health assessment of equipment has become a hot spot in recent years. It is critical to bridge the gap between real-time factory data and health status evaluation, which helps decide appropriate maintenance time by quantitative fault-early warning. For this purpose, this paper proposes a framework to realize real-time equipment health management. The framework begins with principal component analysis (PCA) for feature reduction and support vector data description (SVDD) method for identifying abnormal observations. To promote the computational efficiency of the static health assessment model, an improved incremental learning SVDD method based on KKT (Karush-Kuhn-Tucker) condition (KISVDD) is proposed. Then health degree (HD) is defined derived from deviation degree (DD) based on Euclidean distance. Subsequently, a fault-early warning threshold setting method based on sliding window is established to realize quantitative maintenance time prediction. Thereafter, the proposed scheme is compared with different types of algorithms in a case study to demonstrate the effectiveness of the proposed model using actual production data. The results show that the proposed model outperforms traditional ones in accuracy and computational efficiency.