Real-time Identification For Hybrid Testing Method Based On Forward Analysis

Hybrid Simulation Method is a kind of method that isolates the significant sections out of the structure as test substructur e and then conducts the experiment in the labs. O ther sections are regarded as numerical substructure calculated by computers. It’s a method that can reduce the test scale and increase the efficiency. With the enlargement and complexity of the structure, t he scale of the test substructure tends to be expanded. However, the existing experiment condit ion cannot satisfy the need of the complex test substructure with multi- DOF. Improving the accurate of simulation of the nonlinear structure is an effective solution to simplify the test substructureIn order to make the numerical substructure can reflect the real test substructure’s response, we need to effectively identify the model parameters of the test substructure and update to the numerical substructure. With the consideration of the importance of nodes for the entire structure, this paper takes the unknown parameters of the test substructure nodes as recognized objects, takes advantage of the idea of forward analysis, uses the improvement of the uniform de sign method to establish identification library, and finally real-timely updates the identified parameters to numerical substructure to finish the whole hybrid simulation.Using Matlab-OpenSEES mixed programming method to achieve the real- time identification for hybrid test mentioned above with a large number of numerical simulation. The feasibility and effectiveness of the method in this paper is verified by one to four parameters identification. Through the analys is of results we can come to the conclus io n that this method with the high accuracy and stability is not affected by init ial assumptions and peak acceleration.