Predict with AI, Save …energy, time, revenues!
As a CEO/promoter or asset manager, you spend a lot on machinery maintenance…manpower, spares, and thus revenues. It is usual that critical machinery in the industry, such as motors, turbines, compressors, etc …. need hawk-like attention to see that do not fail or underperform. At times we have redundancy in place to ensure backup.
The maintenance procedures- periodic, scheduled, preventive are the established methods for years. Whether a coil is worn out or not, or oil needs a change or not, these are replaced as per their life and usage hours. In the case of condition-based maintenance, the parameters are regularly watched by a dedicated person at periodic intervals…. the record is always subjective! Human judgment leaves room for error.
A study reveals, World over, maintenance burdens are like…
- Operator error is blamed for 18% of unscheduled equipment downtime (Plant Engineering)
- 39% of facilities still use paper records for maintenance reports (Plant Engineering)
- An estimated 20.9% of wasted time for maintenance workers is due to travel to different areas in a factory, with an additional 19.8% as a result of waiting for instructions (Plant Services)
- Scheduled maintenance takes on average about 19 hours a week, with 31% reporting more than 30 hours a week and 14% spending between 20 and 29 hours (Plant Engineering)
- Job growth for industrial machinery mechanics, machinery maintenance workers, and millwrights is projected to be 5% (as fast as average) from 2018-2028 (S. Bureau of Labor Statistics)
- Poor compliance with lockout/Tagout (LOTO) procedures ranked fourth in OSHA’s top offenses for the fiscal year 2019 (IndustryWeek)
- Unplanned downtime costs industrial manufacturers an estimated $50 billion annually (IndustryWeek)
- Aging equipment is the leading cause of unplanned downtime incidents (Plant Engineering)
- The average cost per hour of equipment downtime is $260,000 (Aberdeen)
Predictive Maintenance
Predictive maintenance is one of the top applications which AI-enabled using ML and is expected to have an installed base of 9.8 million within the next five years. This is based on the concept of Condition Monitoring (CM); the process of monitoring a parameter of condition in machinery (vibration, temperature, etc.), to identify a significant change that is indicative of a developing fault. The signatures of all parameters are captured and deviations are alerted using ML. Maintenance intervention is activated only when the need arises…further, the detection happens with a finite change itself, thus allowing time to respond. So far this is a passive monitoring and alerting system.
This can further be linked to actionable features using AI/IoT so that the system becomes nearly autonomous without human intervention. For example, the lubricating oil needs a change due to its density changing. The system would sense and release the fresh oil while discharging the used oil.
Over and above the overriding advantages cited, another major revelation is the analytics that follows the use of CM. Data through analytics are a huge source for further decisions on the machinery……
Think Future, Think Condition Monitoring
Predictive maintenance can reduce machine downtime by 30%-50%, and increase machine life by 20%-40% (McKinsey)