Basic system identification for condition monitoring of turbopumps [rocket engines]

Condition monitoring of rotating machinery using vibration measurements has been improved due to advances in both data acquisition and digital signal processing techniques. For example, one method that has become popular is known as envelope detection, or demodulated resonance analysis. This method requires bandpass filtering of the vibration signal in the vicinity of a structural resonance that occurs in the transmission path between a potential defect and the transducer. This implies that a basic understanding of the structural dynamics associated with the system being monitored is required. It is the purpose of this paper to suggest a basic mathematical model that may aid the analyst in choosing the proper filter to use in the envelope detection technique. Additionally, the paper presents some simple experimental and analytical techniques that provide relevant system identification information for the model. In the course of the paper, the general method of envelope detection is reviewed as it applies to the condition assessment of turbopumps using gearbox vibration measurements. The effectiveness of the method is assessed using data acquired during a ground test of an actual rocket engine that experienced a gear failure. Potential use of the demodulated resonance technique as an effective method in aerospace telemetry applications is suggested.