MIMO Vibration Test Control System: Theories, Algorithms And Implementations

The vibration environment of structures under operational conditions can be replicated more accurately by multiple inputs than traditional single input. This is a new step in the vibration testing area because a more reliable and safe design can be obtained. Focusing on the engineering applications of multi-shaker random vibration and sine vibration tests, theoretical analysis, numerical simulation and experimental research are performed in the dissertation based on multiple-input-multiple-output (MIMO) linear system theory.The approaches to synthesize time domain random signals are studied. An updated method derived from the traditional time domain randomization is set forth. The conventional discrete random phases by registers shifting method are replaced by the introducing continuous uniform distribution phases. The proposed method can produce high quality random signals and is convenient in the software programming. A new way using FIR filters to generate random driving data is put forward. First, the FIR filters are designed in terms of the low triangle matrix from the Cholesky Factorization of the spectral density matrix. Then, the multi-dimension random signals can be obtained through filtering to a series of independent white noise sequences whose coherence matches the test targets. Additionally, a criterion is constructed to examine if a given cross-spectral matrix is rational by its positive definiteness. The criterion can help one to judge whether a specified cross-spectral matrix can be realized.Some basic theories and algorithms are investigated for dual-shaker random vibration test control in the work. A new trade-off algorithm named Proportional RMS Updating is presented. By the algorithm the control spectrum can be converged to the reference smoothly, quickly and stably. A method with the optimum stability to select the principal control matrix by the best condition number is suggested. The difficulty in the control of weighting spectrum among multiple responses is overcome. Together with Proportional RMS Updating algorithm, the above weighting spectrum (maximum, minimum or average) control problem has been solved for the first time.The theoretical algorithms about the multi-shaker sine vibration control system are studied. A method for the amplitudes and phases control of multiple sinusoid responses is put forward and a technique to measure the FRF matrix with the responses under control simultaneously is founded. In the technique the minimum least square solution of vibration equations is used to estimate the FRFs. Thus, the singularity in the measurement under multiple sinusoids exciting is avoided. A necessary condition for the iterative convergence for the control of multivariable system with uncertainty is derived and the theoretical foundation of the above method is set up.The development of the dual-shaker random vibration and multiple sine vibration test control system is accomplished in the research. The control system is developed on an industrial computer with the VXI modules and high precision DAQ boards as hardware platforms for signal acquisition and generation. The whole architecture of the control system is advanced and highly expandable. The control software is programmed by Visual C++ with multithreading technique and it has a friendly user interface and a high ability for the system protection. The speed of signal process is increased greatly and the time of control loop is decreased remarkably due to the using of the SPL (Signal Process Library) in the software coding. The SPL is a specially optimized library for the Pentium series processors by Intel Corporation.A great deal of experiments and application research are completed using the developed system for dual-shaker random vibration and multi-shaker sine vibration tests. Firstly, all theories and algorithms mentioned above are verified and corrected by lots of tests with different small articles in the laboratory. Then, many functions of the system are improved further through a series of experiments with larger engineering structures. Finally, the system is applied successfully to the research on a national defense project.