Traditional vibration analysis does not clearly indicate problems until late in the failure process due to limitations in the accuracy and complexity of processing vibration data. A fundamental problem is that traditional analysis cannot tell if vibrations are due to early stages of damage or are just normal vibrations of a healthy machine. New prognostic technology measures stress-wave energy and then makes a quantitative estimate of friction and shock events over a machine’s life span. This could let operators more accurately predict failures.
Called Stress Wave Energy (SWE) analysis, the new technique differs from traditional vibration analysis in the type of vibrations it monitors. The SWE approach is based on high-frequency, structure-borne sound rather than machine vibrations, so it filters out noise (unwanted signals) generated by the machine’s normal motions. Instead, SWE looks at the amount of friction and level of impacts. Once it detects an impact, its energy content is measured to establish a trend that predicts how the machinery will deteriorate over time.
In one case, an SWE sensor combined with an artificial-intelligence monitor demonstrated accurate diagnostics for a helicopter gearbox. SWE results caused by 10 intentionally set faults led SWE to predict a 100% probability that technicians could find gear or bearing damage after another hour of operation. Further testing on healthy machinery showed that the false-alarm rate was below 0.1% during 1,000 hr of operation. In addition to detecting faults, the software developed for the StressWave system can locate a fault, isolate its cause to either a gear or a bearing, display the percent degradation, and estimate the remaining useful life.
Another demonstration was carried out on airport surface-detection equipment (ASDE). The mechanical drive of an ASDE-3 radar antenna uses a motor with an intermediate shaft linked to a main shaft. While originally designed as intermittent-use radar, the ASDE-3 now works 24/7.