Written by Bryan Rabenau, VP Technology
ONYX InSIght
It may seem surprising to some just how many wind turbines are running without condition-monitoring systems (CMS) for the major components. It is now typical to install CMS in offshore wind turbines from day one. However, many older turbines and most new models are without fitted monitoring systems.
The main reason for this is the high capital expenditure (CAPEX) costs. CMS units with conventional piezoelectric (also known as ICP or IEPE) accelerometers were simply too costly to justify for many owners. Fortunately, in recent years, modern MEMS (Micro Electro-Mechanical Systems) sensor technology has contributed to a dramatic reduction in the cost of CMS hardware, further strengthening the business case for installing a monitoring system.
ONYX InSight, a predictive maintenance and analytics company, has developed a MEMS-based solution and paired it with the company’s widely adopted fleetMONITOR software. In a recent case study, Alistair Warwick, VP at BP Wind Energy, talked about how this disruptive technology changed their return on investment calculation.
“We looked in the past at retrofit but it wasn’t affordable,” he said. “However, ONYX InSight brought a new value proposition to the table and combined with their track-record of quality engineering with BP, it was an easy decision to move ahead with adopting this technology.”
During the 2018 NA Wind Turbine Technical Symposium, in Golden, Colorado, Stanton Peterson, also of BP, noted that for the first 50 turbines fitted, ONYX InSight detected 10 drivetrain issues. This turned into considerable cost saving and BP Wind subsequently installed over 300 additional systems across five wind farms. (Read the full case study here.)
What is important to detect?
Today’s world is all about quality data and wind-turbine reliability is no exception. ONYX InSight maintains a database of every turbine fault it detects for the purpose of tracking performance metrics, assisting in OPEX forecasting, and other useful engineering works.
From this database, a small subset of data was recently analyzed, consisting of 1,000 wind turbines, 10 turbine models, and 14 gearbox types. The resulting distribution of confirmed faults is shown in the “Summary of drivetrain failures” chart above. As can be seen, the prevalent failures are bearings, and particularly main bearings and intermediate stage bearings (although problems do extend from the turbine’s blades to generator).
CMS hardware and software clearly needs to be able to detect all these issues with sufficient lead time to aid in maintenance planning.
This database, gathered from years of monitoring and consulting work, was invaluable in developing ONYX InSight’s affordable ecoCMS concept.
“Given our knowledge on data sampling, bandwidth, environmental conditions and networking requirements, we were able to develop a specification to exactly meet requirements, but always keeping in mind the need to minimize overall cost without compromising quality and reliability,” explains Dr John Coultate, ONYX’s Hardware Product Manager.
He added: “Interestingly, MEMS allowed us to develop a more powerful solution as tri-axial sensors are standard, temperature is embedded in the sensor unit and low frequency performance is outstanding. We can also add a range of oil sensor technology via the auxiliary inputs.”
So, what are MEMS sensors?
MEMS sensors are microscopic devices, typically etched from a layer of silicon and mounted inside an enclosure with integrated electronics.
MEMS were first commercialized in the 1980s but today they are mass-produced devices and feature in a huge number of home electronics and industrial applications. For example, every car leaving the production line today typically has dozens of MEMS sensors including accelerometers, gyroscopes, pressure sensors, flow sensors, and inclinometers. Mobile phones and video games use MEMS accelerometers and gyroscopes.
In fact, it is estimated that over 11 billion MEMS sensors were produced globally in 2018. There are significantly more MEMS sensors than people living in the world today.
Applicable to all wind turbines
Over the last 12 months, ecoCMS has been fitted on over 25 different turbine models, with the highest volume on various GE, Nordex, and Vestas models. Every turbine installation requires a detailed work instruction to be developed (this describes the safe installation process, including sensor locations, cable routing, electrical installation and mounting details).
ONYX InSight has developed off-the-shelf work instructions available for all major turbine types, which significantly speeds up and de-risks the installation process. Depending on the size of the installation and the requirements of the wind farm, staff are either provided by ONYX to perform the installation or training is provided to the client’s technicians.
Out with the old, in with the new
In the late 2010’s, the wind turbine CMS market was relatively young and there were vendors who have since exited the market. Consequently, many turbines in the field have CMS that is no longer supported by the manufacturer. These legacy CMS units are typically based on outdated technology, with limited processing power and with moving parts such as fans and rotating hard-drives that have high failure rates.
Newer technology, such as ecoCMS, solves these issues. ecoCMS is an innovative condition-monitoring system that combines cloud-based software technology with affordable hardware. According to ONYX Insight, many owners have performed widespread removal and replacement of legacy systems — with one customer scrapping over 600 outdated CMS units and completing the final phase of replacement with ecoCMS in 2019.
ecoCMS routinely detects damage on wind-turbine high speed components, such as this example of a high-speed bearing fault on a GE wind turbine. In this case the fault was obvious both above and below 2 kHz in the vibration spectrum
One case-in-point, as Pattern Energy NA recently stated in a press release: “New contracts will see ONYX install condition-monitoring systems on 234 turbines across two wind farms,” driven by a will to decrease their OPEX expenditure. Meanwhile in Europe, Iberwind are using ecoCMS to “upgrade 76 Vestas V90 wind turbines…for monitoring of the drivetrain, rotor balance, and vibration-based life for failing components.”
“It’s a good start, fitting out part of the wind-turbine population that was running in the dark,” said Dr Ashley Crowther, Global VP of ONYX Insight. “Today, we can see every major component failure well in advance of repair using vibration signals, but only if we have the signals. That we could change the value proposition and allow more turbines to have CMS coverage and help the competitiveness of wind through reduced OPEX is very satisfying.”
He added: “One set of 500 turbine installations in the second half of 2018 uncovered over 5% of the turbines with confirmed damage — finding that is coming out of the dark.”
When using ecoCMS with fleetMONITOR, every turbine can be monitored for rotor imbalance problems — whether caused by aerodynamic imbalance such as a pitch tracking issue, or a mass imbalance such as a blade with water ingress.
Myths about MEMS
When new technology is introduced to a market, it is common for people to be resistant to change. Remember digital vs. analogue cameras? Here are a few MEMS myths…
Myth I: MEMS sensors don’t have a high enough bandwidth to cover the high-speed components
The latest MEMS technology delivers 10 kHz frequency bandwidth, which is more than enough for a wind-turbine drivetrain. The “ecoCMS” screenshot above shows a fault that was successfully detected on a turbine’s high-speed shaft bearing by using ecoCMS with MEMS sensors.
Previously, MEMS sensors were unable to measure high enough frequencies for this type of application but that is no longer true today.
Myth II: MEMS sensors fail to detect rotor imbalance or main bearing faults
MEMS technology has the ability to measure down to 0 Hz, making them particularly suitable for low-speed detection compared to ICP accelerometers. Additionally, tri-axial sensors are standard with ecoCMS, measuring vibration simultaneously in X, Y. and Z directions.
An added benefit: if a MEMS sensor was ever to be removed by a technician (e.g. during maintenance or by accident), the gravity force vector will change, providing a form of self-monitoring.
With better low speed dynamic range than ICP accelerometers, ecoCMS’s MEMS sensors offer superior performance for main bearing fault detection. This turbine had a damaged main bearing that was replaced 10 months after first detection.
Myth III: MEMS sensors are new technology so are too risky for wind turbine CMS
MEMS accelerometers have been commercially available since the 1980s and proven in countless applications as diverse as mobile phones, aircraft navigation systems and hard disk drives, with a global market worth over $53 billion in 2018. As a matter of fact, MEMS sensors have already found applications in the wind industry.
The technology is commonly used in wind turbine safety systems – shutting the turbine down in the event of excessing tower sway.
Filed Under: Featured, News, O&M
