Multivibration Sensors Data Fusion Codec for Mechanical Digital Twin

The state perception of the digital twin (DT) relies on various sensors in the physical body, which generate massive data, and data transmission is an essential intermediate bridge between the physical bodies and the twins. In a mechanical DT system (MDTS), vibration is an important feature that reflects the physical body state, and vibration data generated by mechanical vibration sensors are extremely large, which results in data transmission latency. Therefore, this study proposes a fusion codec method for reducing data transmission and enhancing data transfer efficiency in MDTS; compared to traditional single-sensor data compression methods, the approach proposed data fusion codec (DFC) achieves an exceptionally low compression ratio (CRO) and superior data reconstruction accuracy. First, the transformed coefficients are classified by the screening principle and different quantization bits are designed to quantize the classified coefficients differently according to the mapping interval values, which initially reduce data bits. Based on the quantized data, the fusion of different vibration sensor data is achieved through bytes fusion to decrease the overall data information entropy. To further decline bytes redundancy, the bit fusion method is proposed. Furthermore, the fused data structure is designed to implement the codec of the fused data on the physical body and the twin body of MDTS, respectively. Finally, the performance-related parametric and comparative experiments are implemented separately based on the DFC, and the experimental results demonstrate effectiveness and superiority.