Accelerated algorithm for BTT identification parameter with GMC sparse regularization

Accurate identification of vibration parameters from blade tip timing (BTT) undersampled signals is essential for rotating blade vibration monitoring. However, the traditional parameter identification method of BTT signal depends on the prior. The existing sparse regularization method underestimates the reconstructed signal amplitude and has low computational efficiency. This paper resorts to an accelerated algorithm for BTT identification parameters based on generalized minimax-concave (GMC) sparse regularization to accurately and quickly identify amplitude and frequency parameters from undersampled signals. For amplitude underestimation, the non-convex GMC penalty is introduced so that the sparsity of the estimation is improved, and the convexity of the cost function is preserved. Moreover, Nesterov's accelerated iterative computation strategy is resorted to rapidly improving the convergence performance of obtaining the global optimum. The simulation results show that by reconstructing the BTT signal, the presented parameter identification algorithm based on accelerated generalized minimax-concave (AGMC) improves the computational rate with the inherited merits of accuracy.