Surviving the data storm using rich data structures at data recording and data warehouse

Abstract

Data collection and analysis for machinery condition monitoring has been completely revolutionized due to the advances of the personal computer (PC). Processors run at GHz speeds, terabyte hard drives at very low cost, and Ethernet networking links systems across the globe. This technology has enabled engineers to perform all types of machinery analysis from thermography to vibration analysis to oil analysis and structural analysis. Engineers are connecting accelerometers, displacement probes, tachometers, cameras, microphones, thermocouples, strain gauges, and a lot more sensors to take the measurements they want. And the result of all this technology is an overwhelming mountain of data that engineers are left to sort through to understand their machinery. This challenge is further complicated by dedicated data collection systems which record just one type of sensory data, such as vibration. Even so, data collection technologies now record data in high fidelity, using 24 bit analog to digital converters, and stream more of it to disk than ever before.