Soon after a wind farm’s commissioning, it is critical to adopt a cost-effective operations and maintenance (O&M) strategy to maximizing the project’s long-term profitability and return on investment. Condition monitoring is a typical O&M tool that helps wind-farm owners and operators track the health of turbine components and related electrical systems. Its purpose is to assess the current condition of turbine assets, predict potential maintenance issues in components, or both.
The advantage of predictive maintenance is that it lets wind operators proactively plan repairs or replacements, and only when needed so as to avoid unnecessary and costly up-tower jobs. One recent study showed that most wind farms still use a reactive maintenance system for turbines, and that wind operators could save millions of dollars by using new preventive-maintenance technologies that identify problems before they result in unplanned downtime.
One way wind operators benefit from predictive maintenance is by first setting alerts to signal when a diagnostic signal crosses a specified set point. Then, for example, if an operator were alerted to a temperature increase in a turbine’s generator, ideally a wind tech would quickly fix the issue before it becomes a serious problem.
However, a recent study cited in Wind Energy O&M Report 2017, New Energy Update compared the performance of predictive and condition-based O&M strategies, and found that a predictive approach was not always the most effective. The report used a scoring strategy to measure the performance of different sensor configurations under various key component failure scenarios. The research examined 3-MW turbines on a 630-MW wind farm, and 2-MW turbines on a 420-MW capacity wind farm.
Results showed that a condition-based monitoring strategy using all sensors (except oil sensors) was the optimal O&M strategy for the 630-MW wind farm, but the preventative approach was more effective for the smaller, 420-MW project. Overall, a preventive strategy was more costly, and particularly under a gearbox-failure scenario.
So how does a wind operator effectively choose a condition-monitoring system (CMS) for a wind project? According to David Clark, President, CMS Wind, it is an important question. The answer should be wind-farm specific. “Part of the problem is the erroneous assumption that all condition-monitoring systems for wind turbines are the same,” he says. “That’s like saying all cars are the same.”
Although condition monitoring is now typically considered a must in the industry, the sheer number of different systems available has led to problems. “Not all systems are equal,” says Clark. “And buying an ineffective condition-monitoring system is costly, it can be damaging and has contributed to a slower acceptance of CMS in the industry. It is truly a buyer-beware scenario.”
Clark says that contrary to the advertising, not all systems are ideal for an application and typically more than one tool is necessary. For example, oil-related sensors and filtering systems are typically focused on gearbox monitoring (which make up about 50% of the drivetrain failures), while vibration-based systems tend to work on the drivetrain as a whole. Conditions and events, such as icing or lightning, also require the correct sensors and properly configured CMS.
“Take sensors, for example. There are only so many signals and measurements one can pull from a sensor. Unfortunately, there are many variables beyond just the sensor and hardware installed in a nacelle,” says Clark. “And if the core CMS system has incorrect hardware, issues go undetected.”
To compound this, when measurements are not correctly established, expect missed detections and false alarms. “It is also important to consider a system’s software capability to store alarm and measurement criteria, and perform analysis for proper preventative maintenance,” he adds.
When assessing potential risks to a CMS program, Clark says there are a few points to consider.
- Incomplete data. This includes a lack of full data and access to report configuration, or simply no one properly reviewing the available information.
- Poor analysts. Incorrect calls and lack of experience in wind has, in some ways, created an illusion that CMS is ineffective.
- Ineffective monitoring. A poor system choice for a specific wind farm, or poor communication between analysts and wind operators.
- Ease of use. Data must be available in real time and in simple, easy-to-understand terms or formats for proper reporting, analysis, and understanding.
- Predictive maintenance. Proper component monitoring can lead to significant cost savings. Case in point: a single onshore event involving a crane is $300,000 on average for a 1.5-MW wind turbine. Avoiding the crane cost by predicting the failure using predictive CMS drops the per-event cost to $12,000 to $15,000.
Ideally, a CMS will also include predictive analysis. “However, just because a wind turbine is optioned with preventive monitoring does not mean that it is the best solution or even a viable one. Do your research,” Clark advises. “Education is invaluable and key to a productive, well-run wind farm.”
This article was part of the 2018 Renewable Energy Guidebook. View the full publication here.