No Predictive maintenance of wind turbines
Inference Labs is helping a global Renewable Energy leader to develop an intelligent framework to enable model-based predictive maintenance
• Manual time spent on work order creation reduced by 300 man hours
• Saved $7 M due to reduced machine down time via predictive maintenance of turbines
Current State
• The client is a leader in renewable energy principally involved in the fabrication and construction of wind farms
• Client’s team currently uses a heuristic model-based framework for enabling predictive maintenance
• The team would like to move towards a fully automated and analytical framework for establishing predictive maintenance across all its wind farms
Gap
• Need to identify the relationship between parameters and turbine faults
• Need to establish a comprehensive escalation matrix for issue severity identification
Future State
Outcome: The team observes decreased downtime due to appropriate predictive maintenance of wind turbines and does not need to constantly monitor the framework
Behavior: The team is able to free up time from manual maintenance of the predictive maintenance framework
Insights: Inference Labs has developed an intelligent and automated framework which identifies turbines at risk of fault, creates relevant work orders and recommends appropriate corrective actions
An ensemble model will enable identification of time and reasons for device failure
Time series prediction of device health and performance parameters will answer when the device would encounter an issue
Appropriate data exploration will yield necessary parameters for model creation, resulting in a robust issue identification framework
Machine learning will be used to identify the severity of an issue based on results of previous models i.e. time to failure, the reason for failure and associated cost
Past resolutions for similar issues will be used to generate the best possible resolution recommendation for an predicted problem