Scribd
Upload
143 views2 pages

Ai (Artificial Intelligence) Techniques For Predictive Maintenance

This document discusses predictive maintenance techniques using artificial intelligence. Predictive maintenance aims to eliminate unplanned shutdowns by using sensors and data analysis to detect anomalies and predict failures. The data is analyzed using techniques like anomaly detection to determine if maintenance is needed. This approach can reduce maintenance costs by 25-30% and equipment downtime by 35-45% according to the US Department of Energy. A key step is anomaly detection which finds observations that do not conform to normal patterns using methods such as clustering and neural networks. The goal is to both detect anomalies and predict remaining useful life to schedule maintenance optimally.

Uploaded by

Greenkings
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
143 views2 pages

Ai (Artificial Intelligence) Techniques For Predictive Maintenance

This document discusses predictive maintenance techniques using artificial intelligence. Predictive maintenance aims to eliminate unplanned shutdowns by using sensors and data analysis to detect anomalies and predict failures. The data is analyzed using techniques like anomaly detection to determine if maintenance is needed. This approach can reduce maintenance costs by 25-30% and equipment downtime by 35-45% according to the US Department of Energy. A key step is anomaly detection which finds observations that do not conform to normal patterns using methods such as clustering and neural networks. The goal is to both detect anomalies and predict remaining useful life to schedule maintenance optimally.

Uploaded by

Greenkings
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 2

AI (ARTIFICIAL INTELLIGENCE) TECHNIQUES FOR PREDICTIVE

MAINTENANCE
(BASED ON THE PRINCIPLES OF ANOMALY DETECTION)

Maintenance
Maintenance involves functional checks, servicing, repairing or replacing of
necessary devices, equipment, machinery, building infrastructure, and supporting utilities in
industrial, business, governmental, and residential installations.

Types of Maintenance
 Reactive Maintenance
• Implemented right after a defect has been detected
• Extremely costly to repair highly damaged parts

 Preventive Maintenance
• Catching and fixing problems before they happen
• Challenge? – When to do maintenance
• Scheduling maintenance - very early, we are wasting machine life that is still
usable, and this adds to our costs

 Predictive Maintenance
• Eliminate unplanned shutdown
• Estimate time to failure of a machine
• Find the optimum time to schedule maintenance

Figure: Comparison of Maintenance Techniques

Predictive Maintenance & Sensors


• Industrial Equipment are constantly observed via sensors
• Sensors are attached to components of the equipment and feed constant, real-time data
to CMMS (Computerized Maintenance Management System) software
• CMMS then interprets this data and warns technicians
• Using AI techniques along with CMMS – We can do following tasks
 Detect the anomalies
 Detect failure patterns
 Provide early warnings
Predictive Maintenance – Advantages
 According to US Department of Energy
 Return on investment is increased by 10 times
 Maintenance costs are reduced by 25% to 30%
 Equipment failure phenomenon is reduced by 70 to 75%
 Equipment downtime is reduced by 35% to 45%
 Output has increased by 20% to 25%
 Replace up to 30% of your Preventive Maintenance tasks
 According to Shell, AI-enhanced predictive maintenance can lead to savings of 20%
or more on maintenance costs for key systems.

Anomaly Detection
• Anomaly detection is the key step for predictive maintenance
• Anomaly detection is the task of finding observations that do not conform to the
normal
• Different Approaches
• Supervised Learning
• Support Vector Machines
• Artificial Neural Networks
• Unsupervised Learning
• Principal Component Analysis
• Clustering
• Semi-Supervised Learning
• Autoencoder

Outcomes
• Detects the anomaly
• Predicts whether there is a possibility of failure in next few time period
• Predicts Remaining Useful Life (RUL) [Remaining useful life (RUL) is the length of
time a machine is likely to operate before it requires repair or replacement.]

Dataset
• Intelligent Maintenance System (IMS) Bearing Dataset also known as NASA PCoE
(Prognostics Centre of Excellence) Bearing Dataset
• Link: https://ti.arc.nasa.gov/tech/dash/groups/pcoe/prognostic-data-repository/

You might also like