Reliability and Early Fault Detection for Control Valves
Oil & Gas / Process Industry
Develop a health indication to determine whether a valve is in good or bad condition
Delivered a health index model and provided early indication of faults in control valves to reduce maintenance and maintenance costs, prevent downtime, and improve operations for valve end users, potentially reducing total control valve lifecycle costs by up to 70% or more
A global control valve manufacturer was looking for a collaborative partner to affirm their established condition-based monitoring approach through data analytics. Condition-based monitoring can often present deployment and system integration challenges and demand skills and domain knowledge beyond the capabilities of OEM personnel. With this in mind, the company wanted an experienced data team to develop a health indication model and determine the condition of the control valves used by their customers, delivering an improved proactive approach to maintenance.
The valve manufacturer provided Predictronics with test-bed data sets and other field data for analysis. Our team built multiple health models and leveraged our combined industrial domain knowledge to group the full stroke motion of each valve profile into several segments and more closely examine aspects of the valve condition. In addition, features were extracted from the data to characterize valve performance with respect to friction, supply pressure, stiffness/slope, position, and other signals. Predictronics then ranked these features and developed a data-driven health index model, enabled by our PDX platform.
With this health index model, the valve manufacturer was able to monitor their valves, determine the root cause of developing faults, and prioritize maintenance for valves in the worst condition. By detecting these faulty valves early and crafting a proactive maintenance strategy, the company was able to lower maintenance costs, reduce overall downtime, and improve operations for valve end users, potentially reducing total lifecycle costs for their control valves by up to 70% or more. This valve health monitoring approach also identified actuator, leak, and friction issues, as well as other degradation problems and faults.