Acceleration response data completion method and device based on EEMD-MultiCNN-LSTM

Abstract

The invention discloses an acceleration response data completion method and device based on EEMD-MultiCNN-LSTM. The method comprises the following steps: collecting the acceleration data of an LNG storage tank, and obtaining an acceleration data sequence matrix; adopting an ensemble empirical mode decomposition algorithm to decompose the acceleration data sequence matrix to obtain an EEMD decomposition data sample; and inputting an EEMD decomposition data sample and historical acceleration data measured by missing data measurement points into an MultiCNN-LSTM neural network model, performing iterative training, optimizing model parameters, collecting real-time acceleration data of the LNG storage tank, inputting the real-time acceleration data into the trained MultiCNN-LSTM neural network model, predicting through the MultiCNN-LSTM neural network model to obtain acceleration prediction data and complementing missing acceleration data by using the acceleration prediction data. The method is realized based on the EEMD algorithm and the MultiCNN-LSTM model, so that the prediction precision of the acceleration prediction data is high, and the acceleration structure response of the LNG storage tank can be accurately evaluated.

Description

technical field [0001] The invention relates to the technical field of LNG storage tank acceleration response data completion, and specifically discloses an acceleration response data completion method and device based on EEMD-MultiCNN-LSTM. Background technique [0002] Acceleration sensors are of great significance for evaluating the dynamic response of LNG storage tank structures. During shaking table experiments, some acceleration sensors may fail or become abnormal due to long-term work, resulting in data loss, and these data are difficult to recover after loss. [0003] At present, there are two main methods for predicting LNG storage tank structure missing sensor data based on artificial intelligence methods. One is the "shallow" machine learning method. The acceleration sensing data is highly nonlinear and non-Gaussian. The "shallow" model has certain limitations in the long-term prediction of the acceleration response, and cannot handle massive amounts of monitoring...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the object of the present invention is to provide a method and device for completing acceleration response data based on EEMD-MultiCNN-LSTM, so as to solve the problem of difficult recovery and completion of acceleration data lost due to acceleration sensor failure or abnormality in the prior art. The problem

Images