.png){kind=icon}
.png){kind=icon}
Return
Microservices-Based Software Architecture for Implementing Digital Twins in Real-Time Industrial Process Control
.png){kind=icon}
OA
AI
S
M
O
DOI：10.3390/pr14162584.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This study proposes a microservices-based software architecture for implementing digital twins (DT) in real-time industrial process control. Unlike monolithic approaches, the proposed architecture decouples acquisition, modeling, prediction, and actuation into independent containerized services orchestrated by Kubernetes. Communication is handled via MQTT for lightweight telemetry and gRPC for low-latency control calls, achieving highly predictable, bounded synchronization latencies below 10 ms under nominal edge conditions. A discrete-time Luenberger observer ensures bidirectional state fidelity between the physical asset and its virtual replica. Experimental validation on an edge cluster with TSN compatibility demonstrates elastic scalability, fault isolation, and 100% operational uptime during a continuous 24 h stress test, validating the architectural mechanisms designed to achieve high availability in production environments. The architecture bridges the gap between passive DT monitoring and active closed-loop control, enabling predictive interventions and real-time adaptability.
Keywords:
digital twin
microservices
real-time control
Kubernetes
gRPC
Luenberger observer
Industry 4.0
edge computing
Journal
IF:
2.8
Papers:
6.6K
Citations:
3.7W
Organization
Cited Papers