5 Predictive Maintenance Program Implementation Pitfalls
2020-01-06
Case Study: Pro-active decision-making
2020-01-09
Implementation of Predictive Maintenance using Time-in-State® - Case Study
USE CASE: PREDICTIVE MAINTENANCE
What is Time-in-State®?
Time-in-State® identifies process and equipment issues causing inefficiency. It enables the design of solutions to improve performance.
Time-in-State® extracts information from process data. The methodology integrates process knowledge and experience with process data to deliver team learning and innovation.
The product consists of the following elements:
- Analytics
-
Real-Time Monitoring
-
MESA International endorsed methodology.
Implementation takes the format of a service level agreement. Change management and continuous improvement form an integral part of the implementation.
Time-in-State® is one of the most effective tools to describe process and equipment behavior, communicate process insight, and to establish a common understanding of issues affecting process performance.
The platform translates data into information that is useful to production, maintenance, financial, quality control and management.
Time-in-State® changes the way you think about and use production process data.
Implementing Predictive Maintenance
Equipment performance monitoring aims to deliver early warning of deteriorating performance. For equipment performance monitoring to be useful, the time-lapse between detecting deteriorating performance and functional failure should be enough to implement corrective action. In other words, corrective action should take place before a functional failure affects production.
Without early warning of deteriorating performance, the operation runs the risk of incurring losses as a result of unplanned downtime / process interruption.
In the real world, equipment will be at different stages of their life cycle, feed material variability will cause variance, operator decision-making will introduce changes, and or the configuration changes to the process may have affected equipment behavioural characteristics. These issues requires a definition of a baseline to assess the current state of equipment.
Time-in-State® derives a performance baseline by combining process data, equipment data, and experiential knowledge of the process and equipment application. This baseline is referred to as the Optimum Operating Envelope (or OOE) of the equipment.
Time-in-State® takes process behavior/conditions into account when deriving a baseline to describe equipment performance.
ELECTRICAL MOTOR
Background
The milling circuit of a concentrator is typically designed as the bottleneck as a result of the high capital cost associated with it. Unplanned downtime has a major impact on plant performance.
The circuit described in this case study uses a primary and secondary mill. Both mills make use of dual drives and the electrical motor specification is the same for both mills.
This installation experienced a high frequency of mill motor failures – the symptoms were observed as slipring flashing. Replacing a motor following a slipring flash resulted in 20 hours production loss.
Slipring flashes occurred, what appeared to be, at random. Available measurements failed to detect the condition in time, and it was therefore not possible use the control system for equipment protection.
To find a solution to the problem as quickly as possible while minimizing losses, the following steps were executed:
-
Utilize data analysis to identify the root cause of failure
-
Find a method to generate an early warning of deteriorating performance to prevent slipring flashing
-
Implement a solution to mitigate risk.
Data Analysis
-
Time-in-State® Analytics evaluated available measurements and experiential knowledge to determine if a specific data pattern preceded a failure.
Analysis failed to extract patterns from the data – this confirmed that additional measurements are required to characterize the failure mode.
-
With the assistance of the OEM, current transformers were added to measure the phase rotor current. Measurements were also added to monitor the slipring surface temperature and humidity inside the slipring cage.
-
Time-in-State® Analytics extracted information indicating that electrical current instability on one or more phase precedes a slipring flash. Instability developed 8 to 12 hours prior to a slipring flash – see Appendix A & B. The data pattern distinguished between variability introduced as a result of process changes versus equipment behavioral anomalies.
-
After identifying the data pattern preceding a slipring flash, it was possible to derivea baseline to evaluate the behavior of each motor.
-
Time-in-State® Real-Time compares control system data against the equipment performance baseline.
Real-Time Data Interpretation
The real-time monitoring solutions delivered two results:
A. Early warning of potential failure: Stopping the plant when detecting an anomaly prevented damage to the sliprings.
Early warning of potential failure averts 200 hours’ production loss per annum.
B. Stopping the plant prior to failure enabled evaluation of the mechanical properties of the brush gear.
Root Cause
Physical inspection of the brush gear revealed dimensional changes to the brushes. Dimension changes prevented free movement of the brush in the brush holder causing poor slipring contact. Poor slipring contact resulted in a higher current density on the remaining functional brushes (each phase uses eight brushes). Operating at higher current density caused the brushes to wear rapidly and ultimately failed.
Appendix B illustrates the use of temperature and humidity sensors. The assessment of the absolute humidity prompted a change to the brush specification. The failure frequency reduced after installing the new specification brushes.
Data analysis quantified the impact of seasonal humidity variation.
Multiple factors were assessed to determine the root cause of slipring flashing. Using data analysis provided an opportunity to sift through anomalies generated by lightning, power supply issues, and operational conditions.
Predictive Maintenance
The project team requires a thorough understanding the predictive maintenance application.
Time-in-State® delivers insight by integrating data and knowledge, stimulates innovation and team learning.
Process and equipment data must be translated into actionable information.
Time-in-State® extracts data patterns that associate with specific process and equipment conditions. Real-time identification of patterns translates data into actionable information.
Timely identification of abnormal equipment behavior is vital to a successful predictive maintenance application.
This Electrical Motor Predictive Maintenance solution illustrates how:
- Effective use of data provides an opportunity to gain insight into the issues affecting performance.
- Provide early warning of potential failure –allowing implementation of corrective actions before failure
Appendix A: Rotor Phase Current Model
The Time-in-State® report illustrates the sequence of events prior to the slipring flash that occurred on 2 Jan 2018 at 06:17 (B).
Position A indicates an anomaly on the Rotor’s Red Phase at 16:44 on 1Jan 2018.
The Risk Index (top graph) increases, exceeding the upper limit, around 00:30 on 2 Jan 2018 confirming substantial deteriorating performance.
Appendix B: Slipring Cage Model
Deteriorating performance is confirmed by the Time-in-State® model monitoring slipring surface temperature, slipring cage humidity and air temperature.
Position A shows that the slipring surface temperature and air temperature changes at 16:44 on 1 Jan 2018.
