Sameer Vittal, PhD, Director of Data & Analytics, General Electric
Mark Sporer, Technical Leader, General Electric
Abstract
In this paper, we provide case studies that show how consumer research techniques in JMP Pro can be applied to industrial sensor and maintenance data. This helps guide the development of industrial asset management strategies that are tuned to customer preferences and value models, in addition to traditional engineering inputs. Traditional reliability-centered maintenance systems start with reliability and failure modes analysis, then progress through a series of trade studies where sensors, anomaly detection and decision support models are selected to reduce unplanned failures. Inputs from customer surveys, marketing and economic data are often not used, mainly due to a lack of knowledge of consumer analytics techniques in the engineering community. With JMP Pro, engineers now have access to techniques like latent class analysis, text mining, multiple correspondence analysis, text mining topic analysis, etc. These methods work with traditional reliability and data mining methods, and are used to segment industrial assets using both their operational (sensor) and maintenance data. This drives additional insight where maintenance strategies can be fine-tuned to these industrial segments, resulting in better system usage and increased business value. This is illustrated via two case studies. The first case study deals with usage-based segmentation analysis of coal-fired power plants, and the second study is based on reliability-based segmentation of large wind turbine fleets. Finally, we show how these methods can be used across other manufacturing or process industries.