The engineers at Sandia National Labs are collecting SCADA data from some 900 wind turbines and plan on slicing and dicing it to better understand how the national fleet is performing now and where to focus future improvement activities. At the NREL’s recent Gearbox Reliability Collaborative, Sandia’s Alistair Ogilvie spoke on the value of wind-plant data and what wind companies can do to make it more useful.
Ogilvie works with the Continuous Reliability Enhancement for Wind project, or CREW. “This is a national benchmark that will give wind-plant operators and anyone in the industry, something to compare their turbines to. Its focus is on the operations and reliability performance of the fleet as a whole, not just one manufacturer, operator, or region,” he says. CREW analysts do this by creating a model of each plant and then combining them into the national fleet model. This is to weight the data fairly because the wind farms have varying numbers of turbines installed and have been contributing data for varying lengths of time. This weighting approach ensures that the one plant does not unduly influence the results.
Wind plants are usually located far away, so communication can be an issue. It’s important to be aware that there will be something missing, something unknown. The approach we chose was not to make assumptions about what was happening during these periods of missing data, so we remove the unknown, and do analysis on the known data only,” he says.
That is why companies sould invest in a better communication infrastructure. The more complete the data, the better their decision making. “The CREW project encourages using the full data picture for decisions. The full data picture is made up of the multiple data sources within a company, like SCADA, finance, parts inventories, and work orders,” says Ogilvie.
To illustrate the point, consider a turbine shut down for a month by fault X. For that period, no data comes in. From a SCADA point of view, the fault caused the downtime, but what was done to put the machine back in service is unknown. It could be a gearbox failure or rotor replacement. To fill the omission, Ogilvie encourages more detail from data sources, which comes on work orders with details about the technician’s activities. “If you are not asking for more detailed work-order data, you are missing a huge piece of information.”
Ogilvie says the term “reset” comes up a lot on work-order data. “For example, on a work order that took the technician two hours to fix, he might only write ‘reset’. Surely he did something else. You want to know what he did.”
This is tribal knowledge, he adds. “Talk to any technician in the field and they know exactly what their turbine was doing when they were called to repair it. They are quite knowledgeable regarding each turbine at their plant.
Now consider what happens when that technician decides to leave the company, it loses that knowledge.” One of the approaches to preventing this loss is storing details about the technician’s activities in a computer maintenance management system (CMMS). However, there can be resistance by the technicians to supplying the additional reporting requirements.
“One approach discussed a lot in other industries is to get the technicians to understand why they do something, what the value is to them, the plant, and the company. Once you get that, you have them onboard and now you have the additional and appropriate data source for your decision making,” says Ogilvie.
Sandia’s CREW project is funded by the Department of Energy and partners include Strategic Power Systems, Inc., EDF Renewable Energy (formerly enXco Service Corporation), Shell WindEnergy Inc., Xcel Energy, and Wind Capital Group. WPE
Filed Under: News, O&M