Characterizing System Health Using Modal Analysis

In support of the condition-based maintenance (CBM) philosophy, a theoretical framework and algorithmic methodology for obtaining useful diagnostic and prognostic data from small-scale electromechanical systems is developed. The methods are based on vibration and modal analyses of the physical components. To illustrate the concept of the derived process, an example circuit card is considered. Models are created using finite-element analysis (FEA) techniques and analyzed to determine fundamental mode shapes and vibration frequencies. Simulations are initially conducted on an unadulterated benchmark model. Various fault conditions and cracks that are typical of the modeled circuit card are inserted. Additionally, cracks of increasing length are introduced to simulate dynamic crack growth, representing a deteriorating system. The simulation results yielded vibration modes characteristic of undamaged, faulty, and deteriorating systems. The altered natural frequency response signatures derived from each test signal vector for systems with existing faults and under progressive cracking are compared with the normal operation benchmark signature. An
O
(
N
middotlog
N
) correlation technique is utilized for discrimination. Application of the developed techniques proved useful in characterizing system health by identifying both faulty and deteriorating conditions of the example component.