Downtime, operations and maintenance remain substantial components of wind energy costs, limiting locations where wind power can compete. Sophisticated remote monitoring systems can increase reliability; however, required investments prevent their widespread use with distributed wind. Maintenance costs often reach 50% or more over small wind turbines 20-year lifetimes, and poorly operating equipment can erode overall consumer confidence in wind energy. The 2016 Sustainable Manufacturing, Advanced Research & Technology (SMART) Wind Roadmap identified low?cost predictive health monitoring as a top priority to address this problem. Commercializing such technology will provide a feedback loop on field performance and help significantly reduce the cost of U.S. wind energy nationwide. Our goal is to develop a flexible, inexpensive health monitor and control system for wind turbines of many sizes with a base material manufacturing production price targeted at $100. Expected to streamline maintenance and provide feedback on field performance, this concept is embraced by leading manufacturers as a key cost-cutting measure needed to improve distributed wind American competitiveness. Internet of Things (IoT) devices, wireless communication, and cloud storage will be incorporated, primarily during Phases II and III, with mobile device-supported user interfaces. The research combines sensing and interpreting operational parameters to assess wind turbines health to: (1) Prevent failures, by providing advance warnings and on-site protection; (2) Reduce maintenance, by minimizing unscheduled maintenance and preventing expensive part failures through timely minor repairs; (3) Reduce life-cycle costs, by decreasing inspection costs, customer service demands, and inventory; and (4) Reduce manufacturing and design costs due to over-engineering of components, by providing quantitative feedback from wind turbines in the field. Phase I of the project will: (1) Use vibration spectrum recording as the primary measurement technique, identify and rank critical parameters for predicting wind turbine failures, and determine which faults vibration sensors can predict; (2) Identify the most promising algorithms and methods for detecting trends, an approach common in the gas turbine and reciprocating engines industry; (3) Build hardware prototypes to support the necessary sensors and process their data; (4) Carry out tests on existing tilt-up research wind turbines, for quick results turnaround; and (5) Review business risks and manufacturer acceptance to validate feasibility of the concept. Multiple iterations of sensor configurations and placements will be required to determine the most economically optimum setup, tested on operating wind turbines with simulated defects. The systems commercialization (prime focus of Phases II and III) will be achieved by working with leading American wind turbine manufacturers, tailoring units to their specific needs.
