How Oil Analysis changed the wind industry

I was asked the other day by a friend after a recent government policy briefing if I’m worried about the the country going greener as that will be less oil to test. I laughed and told him even as we move towards a greener future the requirements for machinery to be lubricated and analysis of these lubricants is not something that will just cease to exist. Indeed there are many biodegradable and renewable alternatives and some of the leading manufacturers of these products have used us both in field trials and regular condition monitoring programmes. This brought to my attention that many people simply don’t appreciate the importance of lube oil analysis in the renewables sector. Take your large wind turbines you see up and down the country. They are all filled with oil and grease to lubricate the large loads on the turbine. Many don’t even realise there are thousands of litres of oil and grease in each of these turbines and analysis is a critical part of keeping these machines reliable.

Why do wind farms need oil and grease analysis especially?

Wind farms are a perfect candidate for lube analysis. These being very expensive pieces of machinery where if one catastrophically failed it could would be eye-wateringly costly to repair, the downtime costs could be in the millions and the health and safety aspect of a catastrophic failure would be the stuff of national news headlines.

To top it off these basically run on a constant stop start operation with ever changing loads, potentially stationary for weeks and then have to be at full capacity in a matter of minutes.

Imagine being an athlete doing no training before the olympics and then having to suddenly with maybe 2 minutes warning having to compete in a final. This is what the average wind turbine gearbox goes through all the time. It’s an engineering masterpiece these things turn in the first place. In the very early days of wind, oil analysis was not a big area, but as government subsidies came in and the farms expanded we started seeing more recognition of the value of oil analysis to keep these assets running.

2011 Oil Analysis in turbines picks up

It was 2011 when I first started seeing oil analysis on wind turbines regularly. Sure, samples had been taken before then, but the industry seemed to have an uptick around this time in how much sampling was being used. There was real buy in from OEMs, service providers and farm owners. It was considered a win win for all.

“Wind turbine OEMs could be much much greener with their Lubrication strategy by using Oil Analysis better”

Adam – LearnOilAnalysis.com founder

In the years that followed I have worked with many wind turbine companies and in the various roles I have had I could make whole journeys across the country and most the turbines I saw whether on land or sea were being analysed by one of my diagnosticians. Indeed I have even been consulted in troubleshooting known product issues regarding premature wear and lubrication issues as well as insurance claims for ineffective maintenance and repairs. With all this experience I can say one thing about oil analysis in the wind industry. That is wind farm usage of oil analysis and its data is actually quite poor. In fact I think a lot of them could be a lot greener using the oil analysis data to avoid unnecessary oil waste. This may shock you as do not think of oil waste in the wind industry. Instead we think of a wind turbine being the epitome of green.

Traditionally the wind industry could have done better

I won’t mention names but one OEM used to publish diagnostic advice for oil analysis to aid labs give a consistent feedback on their reports globally. Of the few pages of scenarios given nearly every single one had a recommendation of change oil. Naturally some I would consider valid reasons to drain the oil, but others it seemed very wasteful when there were much more effective solutions that didn’t involve a very expensive and wasteful change. Indeed some of the alternative options were already in place such as offline filtration, but the recommendation was still change oil when contaminated for instance. Although this approach has improved in the last decade there are still practices that could be improved.

Oil Analysis providers could do better too.

Before you think I am attributing blame all in one direction, we need to take responsibility for the problem in the lab industry too. It simply hadn’t adapted to the needs of the wind industry. For example, the advice given on sampling and maintenance was still on a ground based mentality.

Take your average sampling instruction to fill the bottle, empty and refill or to run off some oil before sampling. In the average factory this is fairly easy. Even on ships it’s again fairly easy even at sea, but this is not something that’s easy up in the clouds. If you put those clouds above a choppy sea below it’s even harder.

Another example is the use of trending. Oil analysis traditionally requires at least 4 samples to establish a trend (the exception being LubeWear. I will cover this later). This means the common advice from a lab will be to take a resample to monitor if any issues are found. Now having known quite a few service engineers in the wind industry, I know taking a resample is not an easy thing. Apart from the travel to remote areas to do this maybe several hours drive away, it also needs to be ideal weather conditions to sample.

In this case perhaps ⅓ to ½ the days in the year it is not possible to take a sample. In fact at the local offshore wind farm the operator typically has 100+ bad weather days a year where it not possible to visit the turbines to do any form of maintenance. This means getting a trend of 4 samples or even 2 samples can be immensely difficult in any short period of time. Typically sampling was done 3 to 6 monthly for this reason, but this has migrated to 6 month to annually for others to reduce costs of physically visiting the turbine, meaning you could be 4 years before you could confirm a problem. On top of this ordinary oil analysis catches only a small fraction of the wear in the system meaning failures were being missed. This is clearly not good enough. As a result the approach went to a fear based one.

Fear of oil analysis

I am not talking about the taking an oil sample very high up, which I have a lot of respect for those that can do that; as the height is too high for me. I am talking about using oil analysis defensively rather than proactively. To elaborate, one of these is alarm limits for wear and contamination. To start with, I do get where the OEMs are coming from on this though. Put yourself in their shoes. Imagine you made a very expensive machine and someone asked you how much dirt and water can I put in the gearbox and you keep the warranty intact. Your natural answer is none, and that is how the limits have been set. The fear of failure was so great that limits were set ridiculously low so that even changes within the noise and error of certain tests were being treated as the dividing lines between passes and fails.

It’s not just wind OEMs

It is not just wind OEMs that have this approach. I know big engine manufacturer who used hours based limits have a similar issue. The problem is that if you take an oil only having done 10 hours it fails on every wear parameter as e.g. 10ppm per 100 hours means 2ppm at 10 hours will be a fail, when in fact that is nothing in wear material. Another is a big hydraulic components manufacturer who has limits for cleanliness tighter than those used in international space agencies equipment. Clearly something that is totally unrealistic for the average workshop. Hence, I have always been a bit of a rebel and never blindly used any OEM limits where clearly overly alarming. Instead I have used a more pragmatic approach. Although sadly, I still do see labs who blindly follow these types of limits meaning the value of the analysis is lost if everything is alarmed. After all how likely is it that 90% of your machines have issues? It’s more likely the limits are wrong. This approach has meant people have looked at alternatives to lube analysis.

Industry 4.0 oil sensors is not the answer

Historically oil sensors have literally been a graveyard of failed ideas. I have seen pretty much every sensor out there, even done field trials for the manufacturers of a few and for the most part they all are limited by one or both of two issues: cost and breadth of tests.

In terms of cost you can get an oil sensor for as little as £50 with an ebay search right up to around £10k for some of the more deluxe options. They tend to fit into 4 main categories:

  1. Infra red – these look at oil degradation
  2. Conductance – these look at contaminants that conduct electricity
  3. Particle count – measures particles that block light shining through the oil.
  4. Ferrous – measures magnetism of the sample.

All these in principle will help detect issues. After all they are just simplified versions of a few of the lab tests you would have performed on even the most basic lab analysis. The problem is they don’t detect everything and give the false sense of security and false alarms and good readings. Indeed if a problem is found they all recommend using a full lab analysis to investigate further. So at best they are a screening test only. Even some of the larger systems don’t do it all and they are not cheap either. So you end up with an inferior analysis that you can get a lot of samples done for the same price. The only real advantage of sensor based analysis is the live data, so this works well if combined with oil analysis so you get the great constant stream of low level diagnostic data combined with the really in depth less regular oil analysis from a lab.

The future of lube analysis for wind turbines

I have spoken a lot about how oil analysis used to be and how although advances in sensor technology do improve the industry I ultimately don’t believe they can for the foreseeable be there whole solution. So what is the future of oil analysis for wind? Well this may seem a bit cliche, but the future is actually already here. Advances in oil analysis including LubeWear analysis which can detect abnormal wear earlier than ever before have truly revolutionised the way oil analysis is performed. When it comes to grease analysis this takes a further step up as the ordinary methods of ICP, RDE and XRF for monitoring wear material in grease are so poor compared to LubeWear that I truly wonder why customers bothered with the old style analysis in the past. To bring in mind how the modern analysis has changed things and is detecting faults earlier than ever before I introduce a couple of case studies in the wind sector.

img 3857 726x1024 How Oil Analysis changed the wind industry
In 2019 we were invited to guest write for the wind energy network magazine about the invention of LubeWear and how it was changing the industry. Above is a brief extract of the condition monitoring piece in the issue. To find out more about the wind energy network magazine you can read it here.

Case study 1 – Grease comparison between front and rear bearings

During a trial with a new potential customer (a wind turbine farm owner) the customer took grease samples on the same day from front and rear bearings and sent them to ourselves and the bearing manufacturers current independent lab.

On 12 of the 15 turbines the findings were nearly the same between front and back, but on three of the turbines LubeWear analysis found a significant skew 10 fold towards more abnormal wear in the rear bearing than the front bearing. The turbine manufacturers current lab showed only a ratio of 1.2 which in traditional grease analysis is within the ratio of noise of the test. Equally, the customer also used regular vibration monitoring which is a good complementary tool to oil analysis. Although vibration is not as good at early detection it is very good at localising the fault type. The vibration analysis showed nothing abnormal in any of these samples.

At this point all the evidence would point to the current lab being right and our lab being wrong. However, we explained to the client our findings and did some further analysis to identify the main wear mode appeared to be overloading or imbalance in the loading with very high temperature wear occurring. The customer discussed for 3 months with the bearing manufacturer who said no action was required based on their choice of lab’s analysis. At this point thankfully there was a scheduled major service overhaul of the turbine which the bearing manufacturer took the opportunity to fully inspect the system of one of the turbines we had flagged. A major fault was identified and components were replaced. Over the next 12 months the other two were also inspected by the bearing manufacturer and identified to have the same issue. In fact one of these two the manufacturer estimated there would have been >95% chance of failure within the next 6 months. It is worth noting traditional grease analysis did not detect any of these issues and the first abnormal vibration readings were not detected until 7 months after LubeWear analysis detected the fault.

Case study 2 – Gearbox no false alarms

The previous case study showed how LubeWear helps detect early failures, this one shows how LubeWear prevents false alarms. In this instance, the limits for an OEM were 50ppm for iron as the point to change oil. The last 3 results on the normal ASTM D5185 ICP were 42, 47 and 51 and the current result was 55. Hence there were two consecutive results above the OEM threshold. Ordinarily the increasing wear levels would be an indicator to change the oil. However the the LubeWear analysis showed the for the last 2 samples (at which point we had won the contract) 52 and 55. This meant because both the normal ICP and LubeWear were very close together the wear was normal and likely just related to length of time in use. Our lab recommended simply continuing to monitor and the customer was able to continue for a further 3 samples before an oil change was required in line with the next scheduled overhaul. The inspection during the overhaul found no issues and this meant no unplanned downtime was required.

So what next?

Condition monitoring is in a new era with oil analysis able to predict failures earlier than ever before, especially with LubeWear. What’s more, not only is the analysis more sensitive and able to detect wear earlier than ever before, but it can also reduce false positives and the problem of over-maintenance. If you are analysing your oil already in the wind sector or any other industry, or have never done oil analysis before you should seriously consider testing the condition today with LubeWear, the future of condition monitoring. To find out more click the blue icon at the bottom right of this screen to get in touch to let us help advice you on the best approach to monitoring your machinery health.