Temporal Alignment of Multi-Dimensional Asynchronous Data for Digital Twin Evolutionary Analysis

Digital twin technology is widely used in smart industries and other fields. However, analyzing its evolution in complex scenarios faces challenges in collecting and merging data, particularly in achieving precise data time alignment. Existing methods often struggle with diverse data types and frequencies. This paper introduces a new method for aligning multi-dimensional asynchronous data in digital twin evolution analysis. This approach ensures time alignment and frequency consistency through techniques such as filling null values, intelligent interpolation, and resampling. Key contributions include BiLSTM-based interpolation and flexible adjustment of data granularity. These advancements effectively address the challenge of data time alignment in digital twin evolution analysis, thereby improving data quality and alignment accuracy. The proposed method supports reliable data processing for tasks such as anomaly detection, accelerated evolution, and decision-making. Experimental results demonstrate its capability to recover data with severely damaged features, thereby advancing digital twin evolution analysis in complex scenarios.