Author – Charlotte van Oirsouw, TNO

The impact of data driven innovations on our society and economy is highly dependent on the framework conditions that are set by the regulator. So far, our blogs have provided a bird’s-eye view on the regulatory landscape from legal scholars. Therefore, to get a more accurate picture, this blog will provide insights on the challenges and recommendations for the big data landscape stemming from a commercial context.

We spoke with Bruce Hall (CEO) and Samuel Wharton (Product Owner) from ONYX InSight about data access. ONYX InSight provides global predictive maintenance analytics, mainly in the wind power industry, by analyzing data that comes from wind turbines. Predictive maintenance allows for optimized planning of resources and an increased understanding of wind turbine reliability issues. This in turn allows wind farm owner/operators to reduce operating costs and maximizes the long-term performance of their turbines through smarter operations and maintenance (O&M) decisions.


The wind power industry – Data is the oxygen of competition

The wind power industry has innovated through trial and error and increasingly through data. Even though it is a relatively new industry, it is capital-intensive similar to older ones. In terms of market-diversity, currently, 11 manufacturers account for about 80% of the market; a market that is to a large extend data driven and data-dependent.

A wind turbine produces lots of data, which can be used to predict where and when maintenance is needed. Usually, turbines are placed in very remote areas; think of offshore wind farms for example. Scheduling repairs and more routine maintenance is therefore tricky. As an asset owner, you will have to plan any maintenance months in advance. In addition to that, you are likely to miss the wind season if you cannot get your wind-farm repaired in time, which could lead to less turbines being ready for use, thus a potential loss of income. Any insights that would help in planning turbine maintenance is therefore an almost direct gain in the effectiveness of the wind farm: the largest share of the costs of turbine maintenance is unforeseen. So, if O&M decisions can be optimized by doing analytics, this can bring significant cost reductions, up to 17% of total O&M costs.

Wind turbines produce many different types of data. The biggest chunk of data is SCADA (supervisory control and data acquisition). This data concerns operating variables of the turbine, such as wind speed, power output and blade angles. SCADA data is often pre-processed; it gets cleaned before being reported to the owner through averaging or other pre-processing steps. This cleaned version of the SCADA data does not give the same level of insights in to the real-time behavior of the turbine and as such can significantly hinder maintenance analysis. Real-time data streams could provide important indicators to support long-term predictive maintenance, yet for the moment, this data is often inaccessible. Another type of data is CMS (condition monitoring systems) data, which is gathered by sensors that are attached to the rotating machinery in order to collect and record – for instance – vibration and temperature. CMS hardware sometimes comes factory installed. Usually, the asset owner receives an alarm from their CMS when there is something of concern. These alarms can help the owner make good O&M decisions. However, owners usually do not have access to raw data produced by the sensors. Having this access increases the capability to performance analytics.

Summarizing from the abovementioned examples, access to data allows for accurate predictive maintenance analytics, which provides the opportunity the make smart O&M decisions early on – and most of this data for now is inaccessible. Data access is truly the key enabler of cost reductions in the wind power energy sector, which is especially important considering that subsidies for wind power will start disappearing in the coming years. Only data access will allow wind to stay competitive with other energy sources such as gas and oil.


Data access restrictions and hardware fragmentation

About 3 years ago, ONYX InSight found that the access their customers had to necessary turbine data was restricted. The data is usually restricted in three ways: It is pre-processed and packaged into averages (like SCADA data), it is encrypted or it is obstructed and the owner gets no access at all due to technological or contractual barriers. Currently, up to 80% of the CMS systems that are deployed do not give full access to the asset owners, yet data access issues appear in all data streams that leave the turbine. Only full access to raw performance data allows for an analysis that is detailed enough to ensure the needed accuracy in predictive analytics. Raw data is therefore really valuable.

As explained by Onyx, this accessibility is relevant for all sizes of windfarms; whether you own either two, twenty, or two hundred turbines, you will need to do some form of analytics. These analytics are often done by companies that are specialized in doing so. However, if you cannot get access to the necessary data, you will end up being at the mercy of the manufacturer who will then do the analytics for you.

This issue of data access is interesting, because the owner has bought the asset, yet cannot get access to the data it produces. This is not a technical issue, ONYX themselves provide hardware that gives full access to customers and they can still make a good business out of it. The profit does not lie in the data itself, but in offering good analytics.

To exemplify the data-challenge the wind turbine market currently finds itself in, an analogy can be made with the automotive industry. Just like a turbine, a car is only sold once. Maintaining a position on the aftermarket is therefore also interesting for the manufacturer. A while ago, the automotive industry had to switch from parts that were only mechanical to incorporating assets that also gathered data. These new parts locked up all of the data they gathered. As an effect, if you wanted to get your car fixed, you would have to go to a garage of the same brand as your car, because a general mechanic just had no way to deal with these parts due to restricted data access. This allowed manufacturers to gain power over the aftermarket. However, consumers were not satisfied and authorities stepped in. This allowed for more competition and innovation in the aftermarket and took the industry to a better place. Such intervention is also needed for the wind turbine sector.


Lessons learned and recommendations for policy

Currently, the data is being used as a control- or lock-in tool by manufacturers to maintain a dominant position in the aftermarket. As an asset owner, before entering into a contract, it is therefore important to reflect upon what data is needed to effectively run an asset, including how this data is restricted and how this will influence your costs. In addition to that, it is important to distinguish between data related to your asset and IP-protected data, such as the design of the asset. The latter should be kept by the manufacturer – this data is also not necessary to operate an asset. It is important to be aware that data access can be restricted contractually in a very subtle sense. For instance by a contractual clause that grants full access, but prohibits the owner from touching the sensor that records and processes the data. In the end, it is best for the industry if all the data that is needed to run the asset is in the hands of the owner.

The EU has been at the center of the wind power industry, but if this market is not opened up, business will go elsewhere. Manufacturers should therefore present full data access as a competitive advantage, providing it to their customers whereas their competition does not. This will free up the wind power industry, allowing for rapid innovation.

In order to provide useful access, the market will need industry standards. There is no definition on data access yet, so defining what data should be shared with owners in what manner seems like a necessary first step. Currently, this process of standardization is left to the industry, but some kind of standardization body needs to get involved. If a data pool is created, then this should be managed by a group of asset owners consisting of a broad view of user cases.

The issue of data ownership is not quite clear yet, but the access issue is. It is both an issue of competition and regulation. The data should belong to the asset owner, so that access and usage become simpler. Concluding, the new market is not about the data itself, but about the smartness in processing it.


Recommended reading

– About ONYX InSight:
– Article in the Telegraph:
– Onyx Insight Whitepaper ‘Standing in the way of control’:
– ONYX InSight Whitepaper ‘Thinking about tomorrow’: