17:00 - 18:30 Condition-based decision support
This session will describe current and explore application of new methods and sensors for the condition monitoring of wind turbines. Laboratory and field-based technologies will be explored and their operational effectiveness and expected added value quantified.
- Analyse new methods for monitoring turbine health and performance
- Quantify the operational effectiveness of condition-based monitoring
- Explore new sensors for turbine health monitoring
- Quantify the value of prognostics
Lead Session Chair:
Simon Watson, Loughborough University, United Kingdom
Frank Kirschnick (1) F Rebecca Deubler (1)
(1) Cassantec AG, Zurich, Switzerland
Printer friendly version: Print
Presenter's biographyBiographies are supplied directly by presenters at EWEA 2015 and are published here unedited
Moritz von Plate has been leading Cassantec, a firm providing prognostic technologies for reliability and maintenance applications, as CEO since March 2013. Before joining Cassantec, Moritz was Managing Director of Solarlite, an award-winning pioneer in solar thermal power generation. Prior to Solarlite, Moritz gathered 7 years of strategy consulting experience with the Boston Consulting Group (BCG) with a focus on industrial clients in the chemical, steel and construction material industries. Moritz holds an MSc degree in Agricultural Engineering from the University of Bonn and an MBA from Georgetown University. He has received several academic awards, including a Rotary Scholarship.
New Prognostic Solutions: Answering the "when" questions
With the progressing maturity and reliability of wind power technology, and operators’ commitment to asset availability, wind turbines are increasingly equipped with condition monitoring and diagnostic systems. These systems are progressively more powerful, recording and processing growing volumes of vibration, thermal, lubricant, and hydraulic data and drawing inferences on the condition of critical wind turbine components. Such systems allow detection of malfunctions before failure and damage occur, potentially reducing the risk and cost of downtime and lost power output. They may also reduce preventive maintenance and wind turbine insurance costs.
Typically, the earlier the malfunction warning, the greater these benefits are for the wind turbine operator. Monitoring and diagnostic systems, however, are limited in their prognostic horizon. They provide technical insights and detect early anomalies, but cannot forecast them over a substantial time horizon. In other words, they answer the questions of what, where, why, and how defects occur, but not when. For prognostic purposes, most operators still rely on expert gut feel. However, with the advent of “big data,” and increasing volumes of condition data archived by wind turbine operators and manufacturers, novel prognostic technologies are emerging that can answer the when questions.
New prognostic technologies borrow the stochastic process models used in finance and other industries, and are positioned as “killer applications” of industrial digitization, as propagated by industry 4.0, the industrial internet, and the internet of things. Such stochastic process models surpass the applicability and capability of augmented Weibull techniques, which had traditionally been at the forefront of industrial reliability management with regard to asset life cycle prognostics and remaining useful life (RUL) computation. Complemented by best practice techniques from operations research, artificial intelligence, and data mining, novel prognostic solutions drive a fleet-wide automated, distributed learning process allowing detection and prevention of mechanical and electrical defects of wind turbines earlier and with higher accuracy than traditional condition monitoring and diagnostic systems.
Main body of abstract
As extensions of predictive diagnostics, intelligent prognostic functions for wind turbines enable the shift from time- to condition-based maintenance. Specifically, they promise benefits in three areas: (1) minimized downtime via long-term scheduling and scoping of maintenance, (2) maximized remaining useful life (RUL) of wind turbines through intelligent load management and mitigating condition gradients, and (3) optimal fleet management, considering future asset risk profiles.
Several features of legacy technologies provide challenges for novel prognostic solutions. For instance, some wind turbine data may be subject to imperfect sensor calibration, noisy operation, and fluctuating loads, which can obscure important trends in data. Additionally, data may be archived in legacy formats difficult to convert for sophisticated follow-on analyses.
In our presentation, we will discuss the prerequisites for wind turbine condition data to suit longer-term prognostics, which may go beyond established monitoring and diagnostic standards. Still, the additional effort, e.g. for data conditioning and cleansing, yields answers to many tough “when” questions that wind turbine operators face: “When will my drive train bearing trigger an alarm?”, “When will my gearbox fail?”, “When will my unit trip?”, and “When is my last chance for maintenance?". Answering these questions allows optimal asset operation and maintenance planning, including scheduling and scoping.
Additionally, the prudent application of prognostic solutions presents extended requirements for the industry’s asset management professionals: the ability to think in terms of risk, to explicate forecasts, and to consider both in their decisions. Operator experience and manufacturer know-how feed prognostics, but their application necessitates a shift in thinking towards a risk management approach.
Learning objectives include:
1. Novel prognostic applications for existing wind turbine condition data.
2. Requirements for wind turbine condition data used in such applications.
3. Simple criteria to assess the readiness of wind turbine operators for prognostic solutions.
4. Feasible action items to prepare for use of prognostic solutions for wind power asset management.