This invention relates to the field of preload monitoring technology, and more particularly to an online monitoring method for bolt preload in steel pipes. First, an original
digital signal sequence is acquired based on collected electromagnetic
resonance and
acoustic reflection signals. After preprocessing and truncation, a time-series segment set is obtained. A structural kernel mapping coding mechanism is introduced to construct a set of structural features. Then, based on the set of structural features, an asymmetric
topological graph is constructed, the local topological perturbation intensity of the nodes is calculated, and converted into a topological
tension field in the
time domain. Based on the preprocessed
digital signal sequence, the structural kernel features of the preprocessed
digital signal sequence are obtained, the structural perturbation quantization is calculated, and combined with the topological
tension field in the
time domain, the perturbation inversion structural response function is obtained, and the predicted value of the actual preload is calculated. Finally, the predicted value of the actual preload is corrected to obtain the monitoring result. This method solves the technical problem of non-contact, continuous, and high-precision online monitoring of bolt preload inside closed steel pipes.