Anfang März diskutierten Experten im Rahmen eines Panels auf der PV Operations Europe 2020 folgende Fragen:
- Wie wird sich die Rolle von Digitalisierung und Daten in den nächsten fünf bis zehn Jahren entwickeln?
- Welche Herausforderungen gibt es beim Thema effiziente Datengenerierung?
- Wie verändern sich die Bedürfnisse von PV-Monitoring-Kunden?
- Wie senkt die Digitalisierung die Kosten, wenn es um Ihre OPEX geht?
- Was sind die Vorteile der Datenaggregation oder Konsolidierung von Daten über Anlagenportfolios hinweg?
- Was sind die größten Herausforderungen bei der Datenkonsolidierung?
- Insourcing vs. Outsourcing der Datenaggregation?Was sind Best Practices, die bei der Wind-Digitalisierung genutzt werden können?
Hören Sie, was die Experten von Baywa r.e., Alteso, Isotrol und Solytic zu sagen hatten (in Englisch):
Pritil Gunjan (Navigant Research): Morning, everybody. My name is Pritil Gunjan and I am a senior analyst with Navigant Research, now a Guidehouse company. For those who don’t know us, we are a research and consulting firm. We’ve published reports in the energy sector. We have written various thought papers on how the industry has changed over the years. We started off with coal, gas, nuclear and now into renewables. I lead our distributed generation and renewable generation research services. I’m happy to discuss some of our thought papers. A very warm welcome to our colleagues from Baywa – Virgil is probably online – and then we have Inaccess, Isotrol… I would request our panel members to please introduce themselves. May I have Günter?
Günter Maier (Alteso): My name is Günter Maier. I’m from Alteso because there are some changes on the panel and Alteso is a data analytics company for photovoltaic power plants.
Daniel Watson (Isotrol): Hello, my name is Daniel Ramirez Watson. I work for Isotrol and we’re a SCADA monitoring company with more than 38 GW of renewable energy being monitored by our system.
Pritil Gunjan (Navigant Research): Thanks, Daniel.
Daniel Watson (Isotrol): Thank you.
Johannes Burgard (Solytic): Hello, my name is Johannes Burgard. I’m managing director of Solytic. We do hardware-independent monitoring and build a marketplace for decentralised assets.
Pritil Gunjan (Navigant Research): Hello Johannes. Great to have you here. And we also have Virgil joining online.
Virgil Cazacu (BayWa r.e.): Hi, good morning everyone. Thanks for having me. I’m Virgil Cazacu. I’m leading the digital transformation within Baywa r.e., the service side and we are one of the largest O&M providers in Europe and solar was more than 8 GW under management.
Pritil Gunjan (Navigant Research): Welcome, Virgil. So, kicking off with a stat here, because we are a research company. We estimate that the monitoring and control market for both solar and wind is going to cross 80 billion dollars in revenues cumulatively over the next 10 years. That’s a massive number and of course as we all know PV is going to play the biggest role as far as just digitization is concerned. So, with that number in mind how… Let me start over, Johannes probably with you. How would you say digitization and data or the role of digitization and data will play out in the next five or ten years from now?
Johannes Burgard (Solytic): In general, I think we had yesterday already in the discussion. When we look at digitization in general, when we look at PV, we see an enormous evolvement over the last years. 10 years ago we were already adding value if you showed data. If you actually show the information on your screen and you’re not you’re not relying on going on-site and analyzing everything on site but you could do it remotely. This, of course, is already adding a lot of value now and we’re moving forward. The digital technologies, in general, are moving very fast. We are now moving into decentralized innovation so we find a lot of potential to the left, to the right and especially in other industries. So I think it is very crucial for PV that we do not just focus on our own stuff in PV. But as we just had in the presentation, we have to look into different technologies into different sectors to really leverage the technological evolvement over the last years. Also in the next years, that we can really make advancements here. So I believe that looking into different sectors, looking into different industries, leveraging the inventions of other spaces will help us to really move forward. I see a lot of potential in the automation in the analysis. We have a lot of companies now working very heavily on A.I. This is something that of course always relies on the data quality that you have and the granularity. So we need to invest in hardware. We need to invest in the first level of technology so that we can actually build new features and advance in this space.
Pritil Gunjan (Navigant Research): Great, thank you, Johannes. Günter. The role of data: How would you qualify that?
Günter Maier (Alteso): Well, what we hear from our customers and from big players in Europe, in India is that they would like to have solutions, where the experts in the companies, the analysts, can rely already on a result coming from analytics for the purpose that they can then focus on the more sophisticated things and the actions they should take out of that. We hear from big players in Europe that they are spending most of their time with maybe the 10% to 20% problem sites and plants and the 80% of standard plants remain kind of unanalysed and unsupervised. And also we hear, that people are just expecting, that when the team is starting in the morning, there is already a list of up-to-date analytics results that it really can focus on. And I think it is not about, in this case, reducing teams or replacing people by technology. It is rather empowering people by technology that they can focus on things where maybe just people can focus on. And I think it is just… I would expect that the future is bringing tools, which are like weapons for the people in that sense – in the positive way – where they can just use this equipment for doing their work more efficiently. We also get requests from customers that say: “I would like to have a toolbox where I just would like to select when I need an analytics tool.” It’s a bit of analytics on demand. So that you do not even subscribe for something on a permanent basis, you just choose whatever you want, when you need it. So I think this all will go much more lively and much more specific for the specific need.
Pritil Gunjan (Navigant Research): Excellent. That is the new gold, right? Daniel let’s hear from you.
Daniel Watson (Isotrol): We believe that digitalization is going to be key in PV like we were just talking about in wind. We’ve got to gather all that knowledge and bring it into the PV sector. We have to take into account the drop of costs that we’re having throughout the PV industry. Digitalization has to aid and help in this cost reduction. And we also have to take into account the human error that we always have. Digitalization has to help customers to be able to minimize this human error, that we all might have and improve our assets.
Pritil Gunjan (Navigant Research): Excellent. I’m slightly different take from you, Virgil. How would you say the needs of customers have changed – you being an O&M provider – with regard to the role of data?
Virgil Cazacu (BayWa r.e.): I think the needs of the customers also evolved with the evolvement of the technologies. As now you can get more descriptive and predictive approaches in terms of the data and making sense of the data. And then customers are expecting to see more into the future than in the past as it happened a few years ago. So just creating reports about the past months or past year is by far not enough in the current set up. The data and especially the context around the data. Because you can look up into the data, you can find some issues, for example, in some strings. But without context actually you can not get the right decisions and at the end of the day helping the customers getting efficiency into the plants.
Pritil Gunjan (Navigant Research): Now that’s a really excellent point. So the context of data is critical here. Any other challenges you would like to highlight when it comes to generating efficient data?
Virgil Cazacu (BayWa r.e.): So I think the operations teams make decisions based on the data that they received from different data sources: from monitoring providers, SCADA systems, data loggers, inverters, satellite irradiations, weather forecast service, financial systems. So there are lots of data streams, which are now coming into play and actually from our point of view digitalization is not just a specific technology but actually it’s context and it’s an ecosystem which brings together also the business processes, also the change management teams of people. Because even if you have the best A.I., it has to be proved in the context. And also it’s a human factor involved, where especially the engineers which are working in this area, they need to gain trust that relies on that data, which is coming out from a machine learning algorithm. That this can be trusted and is scalable. across the whole portfolio. So the quality of data is super important because that has an impact on the whole chain, which follows that data ingestion.
Pritil Gunjan (Navigant Research): Thank you, Virgil. Johannes is coming to you. Your experience with generating efficient data and some of the challenges you would like to share with us?
Johannes Burgard (Solytic): It’s more collecting. We don’t generate so we are focused on software but what we have to handle there is quite a diverse field off of data. In our case, we focus on decentralized assets and that means that we are collecting data through email, through FTP pushes, through CSV files and also through VPN or whatever so it’s a very diverse field of data and that’s definitely not efficient. But that’s basically the problem that we are trying to solve. Providing one monitoring solution that combines all these things and that’s I think it’s something that we have to tackle. So the lack of standards and the lack of unification of all these protocols is just something that is definitely hindering us in moving forward fast because it’s impossible right now to buy a plug and play product off the shelf. Simply because it doesn’t exist, because it’s not possible to build it. So there’s a lot of potential to really get closer together and build standards so that we move towards efficient data, so we can really leverage all the intelligence that we have to the left and to the right. And then we can move forward there. Right now we are quite some distance away from that.
Pritil Gunjan (Navigant Research): And how does this relate to the needs of your customer changing?
Johannes Burgard (Solytic): Well, as mentioned before, it used to be already adding value, if you can show the data on your desktop. But that’s a long time ago now and now the customer demands they increase. Everybody wants very smart alarms because nobody wants this flood of alarms. Using static parameters for those alarms, for instance, is something that is very useful for large scale assets, but it’s impossible to adjust these for small scale. I think we heard it yesterday as well, about 50% of the assets and the capacity is actually in residential. So who’s taking care of these people. They usually are not engineers like us. They don’t want to take care of their asset. They want someone to tell them when they have to act. Here it’s very important that we move into that direction that we can help people without technical experience, without the background to still handle PV. Because then it becomes scalable and in the big scale. Then we can actually get renewables really going. Because compared to all the other energy resources that we are using today, PV is still very small. Those are the needs that we experience. That people don’t want to bother with the assets. They want it lean, efficient and there everybody is on the same page.
Pritil Gunjan (Navigant Research): Especially in the distributed energy sector. Günter, your take on the issues and challenges of generating efficient data.
Günter Maier (Alteso): In our case, it’s also not generating data, but it does not matter. We have to focus on extracting data for letting the data flow into our engine. What we see is, first of all, a big variety of quality of data worldwide. But it is also worth to say that we see a lot of issues, where data is missing, unreliable, frozen data. Very often also data are basically there, but the matching in the monitoring is simply wrong. So you’re measuring things where you’ll get the results but you are measuring from the wrong equipment part because it’s just differently defined in the monitoring system. We see our role here in the way where we’re getting this data sanitized. A lot of customers, which maybe are big and which have acquired a variety of projects, they cannot really choose what monitoring system is behind. Sometimes there is almost no monitoring behind. Sometimes there’s monitoring behind, but this is so diverse. And these customers, they tell us: “You don’t have to tell us, that we have a mess. We want you to overcome this mess.” This is why a lot of the work we are doing, is some work which is not necessarily seen because it is kind of working with the data until you have a data set where you say: “Now I can start to analyze.” And I believe this will still go on for quite a while because it cannot change or overnight. I believe this is also why we say: “Automation, A.I., M.L. – these are all very powerful things, which are coming.” But at the moment, honestly, we are working on getting the data right for getting analysis and valuable conclusions for the client.
Pritil Gunjan (Navigant Research): Sure, technology enablers. Yes, of course, we need an A.I. and…
Günter Maier (Alteso): Sorry, maybe one last sentence. Very often on the owner’s side or customer side, they are not aware, what is their data situation until we start with the work and can visualise, how much data are missing. Just an example: We did just a tracker optimization analysis for a plant and at the end, it was revealed that 80% of the tracker position data were missing. So you can not even do an analysis or optimization. But the first information is that we can at least equip the client with the information, that you can then go to his equipment supplier and monitoring system supplier and say: “What’s that?”
Pritil Gunjan (Navigant Research): Isotrol perspective, Daniel?
Daniel Watson (Isotrol): I think this is one of the main topics and I’m glad it comes up. Because poor data, bad quality data – it’s a big problem in the industry. Therefore it’s mandatory to have data quality and data cleansing algorithms that help customers with this problem. And the lack of sensoring, it’s also a big problem in the industry. So we’d like to advise customers early in the game about what type of sensors they need to have and how they can monitor their assets correctly. Then the other big problem that we see is that a lot of PV farms don’t have real-time monitoring. So the data acquisition for that is pretty poor. So we highly recommend our customers to have real-time data being gathered, not every 15 minutes like a lot of PV plants are being done right now. So definitely real-time robust algorithms with data quality and cleansing and having a robust system and that acquisition and censoring throughout your PV farm.
Pritil Gunjan (Navigant Research): Thank you, Daniel. So Virgil, coming back to you. You know just taking what Daniel just mentioned: Having these diverse set of quality and quantity of the way data is generated. How would you then apply it to OPEX? Or do you have some case studies to share, where you’ve seen a real cost reduction when it comes to your OPEX?
Virgil Cazacu (BayWa r.e.): Just to underline one thing. Indeed we are talking about very advanced systems like machine learning, A.I. The reality is that we need to focus on reflecting what we are currently doing. That means perfecting performance ratios, the availabilities in order to meet these objectives and at the end of the day the financial returns for each asset. Having said that and related to the OPEX impact of digital tools, I can give you two examples from our side. One is the optimizations of the plants’ on-site related activities, which is actually technicians. They go on-site and they perform the necessary work to the PV plants which relates also to having better customer satisfaction. Because it’s a difference when someone goes on-site with some paper and pen and they do their work or they go with a digital solution, where they can capture all the findings and all the work. For example, for maintenance plants to capture all the necessary documentation and centralize it and then creating reports for different stakeholders in different formats more automatically, that definitely has an impact on the structures and on the OPEX. And zooming out from the plant to the monitoring part, improved decision making based on data analytics. That’s where we are focusing because we want to have a decrease of production cost to more focused, less operations, where the cleaning or the inspections can be done more focused on the pain points of the plants. So instead of just going for big plants just going and driving hectares and checking with drones all plants then we can, based on the data, have targeted inspections. That can bring optimization also in terms of spare parts management.
Pritil Gunjan (Navigant Research): Thanks, Virgil. Johannes, would you like to share some ideas of how it’s impacted the OPEX of your customers?
Johannes Burgard (Solytic): Well, it’s the point I made earlier. The combination of different monitoring solutions, the different data sources that we’re just combining, so our customers are using one monitoring instead of 15, which is something that is very common probably to everyone who is monitoring a couple of hundred megawatts or even in the gigawatt space. Those companies do it themselves. Of course and that’s good. So it’s not a new idea. It’s just something if you’re just a craftsman and you’re handling 500 assets and they’re all very small, you don’t have access to this technology that you build it yourself. So this obviously helps a lot in becoming more efficient on a daily basis. Something that we are working on together with customers right now is to take the next step. If you identify some default or some anomaly, how do you solve it? So going towards preventive maintenance is the standard, mostly correct of maintenance in the residential space, but then when we want to move towards prescriptive maintenance, so predictive for something we can’t do ourselves. So we are looking for partners there. But I think this is something that is very interesting in the future. To provide solutions and not just the information that something is wrong.
Pritil Gunjan (Navigant Research): And that’s again a really interesting point. So when you look at predictive maintenance versus prescriptive like you mentioned: What is the incremental cost to your customer? How do they really measure it? Do they look at it in terms of cost reductions or revenue targets?
Johannes Burgard (Solytic): They don’t yet. And we don’t have it. It’s not that we’ve found the special sauce to create it, yet. For us really the use case right now is providing one monitoring solution instead of 15 and that is reducing a lot of complexity in the day to day business.
Pritil Gunjan (Navigant Research): Sure. We’ll come back to integration. I think it’s a really pertinent topic in this industry. Günter, what’s your experience with OPEX and digitization?
Günter Maier (Alteso): Concerning this topic, we have case studies which are coming from the analytics of course because this is our business. And in 2018 we did a white paper which is called “OPEX reduction coming from analytics”. Everybody can download this from our webpage. But to cut the long story short, we have described three different categories. One category was about the minimization and the reduction of any kind of physical inspection. So that you need not to spend so much time physically inspect something in order to find the issue. You know already the issue when for example you go either to the site or just to the specific stream. So you do not have to search or to test or to measure first in a quite time-consuming way. You know already where to go and you know already where to fix. On that the reduction of time, which is then of course also a bit of an equivalent in cost, is when an O&M is on-site anyway, then a reduction is possible by about 90%. Because instead of searching for it, they just focus on what they should mitigate repair. And they have done also the verification immediately available. When they do something they see if it was successful. And the customer sees it as well, which is important because it was they’re not you know a useless standard O&M activity at the site. You know already if there was a success. When the O&M team is not at the site, when it’s an external provider, then you can reduce or improve the efficiency of this time for a week for example for a site campaign, where the O&M team is spending on site. Instead of five days, for example, you reduce these to free days because you can just save on some time on the things. I’m just repeating myself here. The other two categories where one example was on soiling. When you know that in the 35% of the plant the soiling rate is for example in lost terms 5%, you could focus on these 35% first or only. And this is then improving in a higher leverage the performance of the plant, let’s maybe say up 1-2%. You’re reducing quickly 50% of your soiling losses and you’re spending also cleaning by 60%. The last category was the spare part management. When you know better for this site what is specifically necessary for that site by experience from analyzing the plant, then you can dedicate better the spare parts, which is there at stock. If this is combined with the asset management system, which has already maybe a spare part inventory, then this can be structured very efficiently. On the question you had before about comparison of preventive to predictive or prescriptive maintenance: I think we have also not lost numbers but I think it is quite easy when you say instead of doing an activity every year – standard – you basically extend these two maybe every two years, every three years. Then everybody can calculate, what it can be in terms.
Pritil Gunjan (Navigant Research): Great point, Günter. I think obviously supply chain and minimizing cost across your supply chain as it is integrated with asset management and overall the monitoring system, I think it does have a very natural progression and that’s how I think holistically it should be looked at. So that’s a great point. Daniel, would you have some case studies to share?
Daniel Watson (Isotrol): Yes, for sure. Very good points brought up so far. We think that improving is our duty and a reduction of O&M cost is part of it. So definitely there’s a lot of customers that are spending a huge amount of hours in the offices, trying to figure out what’s going on with their assets. We need to reduce those hours drastically. Then the other point we need to support the people that are working out in the field. There’s a lot of trackers that are not being monitored. There’s a lot of strings that are not being monitored. There’s a lot of soiling problems. There’s a lot of shading problems. There’s different a performance sense within the same solar farms that need to be monitored and we need to put the focus on the spotlight on all these problems so we save a lot of time for field personnel. And then the other item is having a strong CMMS system to be able to have your like you were mentioning your inventory under control, your people under control, your consumables, your spare parts, everything. And a lot of companies are doing this in-house so you can have everything under control. But a lot of companies are also outsourcing this. So by knowing when you have to do these actions you could also say: “OK. Maybe it’s cheaper to do it at this point in time of the year because the market is not so busy at the moment.” So all of these items that we mentioned, have to be taken into account.
Pritil Gunjan (Navigant Research): Great, excellent point. When data aggregation is another contentious issues and it does come up as a key enabler in drawing the best out of your data. So Virgil, how do you view data aggregation or basically consolidating data across your portfolio of assets and what are the benefits you can reap out of it?
Virgil Cazacu (BayWa r.e.): Yes, that’s a great question and this may sound easy to be done, but actually is quite challenging from an O&M perspective. Centralizing the production data, the past, the importing, project spending, the stakeholder management in one place. It’s a very complex work, especially when you are also in an international environment. But the benefits which this in return – and again, it sometimes a tedious work to do it – is actually great, because you can have as we put this 360-degree view of the assets under management and you can see in one place, in one platform, an asset management platform, you can see all the events which relate to certain assets. And again for us, events are not just the monitoring events and the alarms, but also all the communication with the customers or the reports which were created, all the financials around that system. And that’s actually because we were talking previously about the context that actually gives context to all the data points, which can be very diverse as you can see. It gives sense and you can actually then think about how you can offer added value services to the customers.
Pritil Gunjan (Navigant Research): Excellent. Johannes, this is probably something that you deal with on a regular basis, because your focus is on the distributed energy area. And I know you touched upon this, but dealing with aggregation of data across a portfolio, whether it’s a microgrid or whether it’s your C&I customers, what has been your biggest challenge in terms of consolidation? And would you share an example of how you go about solving an issue like that?
Johannes Burgard (Solytic): Well, we’re still working on it, but, it’s still a challenge. And with new customers there come new challenges. So I don’t think that we do something special there or that we are different from all the other big players because this is something everybody has to deal with. So if you have different data gateways or data loggers in place. Whether you have string, trackers or combiner boxes or whatever data you’re connecting, you always have the challenge to bring this to a level that you can analyze it. As mentioned before, getting cleaned data in the first place to properly analyze it is the biggest challenge. This is kind of our core that we process the data in the way that we can actually then visualize it in one way. But I don’t think that we struggle there with things that not everybody, who does it, has experienced.
Pritil Gunjan (Navigant Research): And is there a particular kind of customer group who are more interested in aggregated data?
Johannes Burgard (Solytic): Well, those who want to work with it. So those who prefer to work professionally with it. Of course, they are looking into the next potential. Whatever you developed last year is now status quo and you have to evolve. You have to move forward. But there are still a lot of installers, small scale installers and small scale EPCs, that are just learning. They are starting now to experience PV. I guess at the end of the day the more professional you are, the higher your standards.
Pritil Gunjan (Navigant Research): Thank you. Günter, OPEX and data aggregation. Do you have some thoughts to share?
Günter Maier (Alteso): We standing here a bit on the sideline, but just on the question you also had before about if there is a specific group of customers. Well, I think it is specifically those customers which have large diverse portfolios. For example, customers which are in a very aggressive acquisition mode, which are looking for underperforming assets. Then they have very diverse assets in size. Also, the local monitoring systems are always different. So it’s a big challenge to harmonize all those sets because you have also different starting conditions. I mean Humberto just came in. I think he could probably speak about two hours about that topic. He is here the better person to speak. But what I wanted to say is about OPEX: The harmonization and the consolidation are bringing a bit of an addition on OPEX because you will also pay for this, let’s say central monitoring or asset management system. On the other hand, I think it is clearly reducing the confusion within the company. Because what we all often hear is that people say: “I have two, three different sets of data and they do not know what is right and how it is correlating.” And I think this is a big task, bringing this together so that you have one data set where you think you can rely on. I believe the costs of confusion in the company is hard to quantify. You see just what you are paying, but you are not seeing clearly what you’re then saving. Because unproductive time is a bit difficult to count.
Pritil Gunjan (Navigant Research): Sure. Daniel?
Daniel Watson (Isotrol): Like we were mentioning, data consolidation and aggregation is key, especially with the very large customers that are joining the industry, adding gigawatts of PV per year. So definitely each customer needs to have a solution that is catered for them. Are utilities going to ask for something different compared to an investor for example? But definitely data aggregation platforms are key and they’re going to have a huge role in these years coming, what we’re looking forward to right now. And one of the other main items to mention about this topic is, what we were talking before and we also mentioned: The quality of the data that we receive for these platforms is key. You won’t be able to work properly with them if you don’t have it. And so is saving decision-making time for people working with these assets. Having gigawatts of renewable energy, PV, in your system, you need to have these data aggregation systems in place, so it can work efficiently.
Pritil Gunjan (Navigant Research): Sure. I mean, scale, like all of you mentioned, is one of the key enablers for ensuring that you aggregate your data. But we also spoke to a few asset managers who say that having different assets and different geographies of the world has also created a lot of issue of data aggregation. So this asset manager that I spoke to had massive GW capacity in Europe and in APAC, and they wanted to put that all on the same dashboards. But they were coming from different O&M companies. So there you go. We need, we need a standardized data aggregation consolidation tool. So that brings us to this very interesting topic of models. Would you shed some light on how it’s done – Insourcing vs. outsourcing? And why have you chosen to do it a particular way?
Virgil Cazacu (BayWa r.e.): That’s not a straight forward answer because it depends on what type of activities you want to perform. Our way of working in Baywa is that we do certain activities in-house. Also, we believe that we have lots of knowledge and we have lots of digital tools that we develop based on this knowledge. Always having also an outside influx of ideas and maybe different ways of even the same things that we are doing, had us to iterate and go forward with new services that also the customers are expecting. And now in terms of if there is a real split of digital services, we are working with startups. So our mostly as our VC arm that we have, we work with different startups to also understand and make sense of what type of digital services can be used and actually can bring value. Because it’s a lot of also buzzwords out there and, also a lot of smoke in terms of what all these buzzwords can bring. And actually, when you put them at work you see that it’s maybe not valuable to the customer or what he is willing to pay. But I think a balance at this moment is difficult to clearly define. Doing all the big data in house, because also there are all the constraints from a contextual point of view. Also, there are all the regulations from each country, which have to be taken into account. And then, of course, the complexity. It’s growing by each portfolio that you are managing.
Pritil Gunjan (Navigant Research): Sure. So Johannes, with your customers, how have you seen this model change? If they were doing it in-house earlier and now they have reached out?
Johannes Burgard (Solytic): This insourcing/outsourcing topic is from my perspective more a question regarding focus. So what Virgil was just pointing towards: You have to think about what’s your focus, what’s your strength, what is supposed to be a strength. And data aggregation for sure is something you can work full time on. And if that’s for instance not supposed to be your focus or your strong point in the market, you shouldn’t do it. But if you have the size like Baywa, I guess that’s the point where it becomes also a key strength. So it does make sense to take this in. Our customers, of course mostly small, there are many which tried to do the entire thing themselves for a very long time. And now they approach us because they don’t want to try everything at the same time. The market is evolving. CAPEX is going down, OPEX is rather rising. But the importance of OPEX is becoming more important. If you’re an O&M, you have to manage this. You cannot do everything at the same time. So if your focus is clear, it’s very a straight forward decision whether you want to insource or outsource a certain task.
Pritil Gunjan (Navigant Research): Günter, your take on doing it themselves and outsourcing?
Günter Maier (Alteso): As we have heard yesterday, there is not one best practice fitting all. But I think there are some rules and one thing is that, when it comes to monitoring systems and monitoring data, I believe a rule should be that the owner owns the monitoring system and the data and not the O&M or somebody else because data has a value. And this value of data is increasing. There are some people in the industry which when they are developing energy management systems, they say: “The data and the knowledge for when the kWh is produced, is maybe higher than the kWh produced one day.” On analytics, outsourcing/insourcing: We believe that one of the best solutions for large companies is that they are using highly developed analytics platforms because this is what we are selling. But I believe it is also advisable that they have expert teams in house, which have the core knowledge of the performance and analytics. This is for the big players. I think for the small players or niche players which have maybe just a finance team, there may be a doable solution is that you use an analytic system, but also the analytics provider is adding a team as technical advisors, so that you get the support as well, not just using the product.
Pritil Gunjan (Navigant Research): Daniel?
Daniel Watson (Isotrol): For us at the end being done in-house or outsourcing it, is not that big of a deal. The importance at the end is to support the company, to support your customer and be able to cater to their needs. There are companies that are growing so fast that they can interiorize all these systems or things that they would like to do. And we like to bring to the table the knowledge that we’ve acquired for working with so many customers, to be able to guide customers and learn from them also. So working in-house or out-house, it depends on the company and how it works for them. Definitely, we would like to give this opportunity to customers to learn from us and to be able to start internalizing some of the services that we provide. So we’re open to both solutions. At the end, doing it right, having good quality data and producing good reports, good KPIs, good analytics that help put the focus on where the problems are coming from, is more important than just thinking: “Should we do it in-house or out-house?”
Pritil Gunjan (Navigant Research): Sure, leveraging best practices, right? And that brings us back to how Daniel started this session today. Would you, Daniel, also share, how you see some of the best practices that can be leveraged from the wind digitization journey?
Daniel Watson (Isotrol): I think, we would all agree that wind is a step ahead of PV. Like we were talking before like we’ve talked in so many sessions: We need to bring that sensoring that we have in wind into PV. When you go and buy a wind turbine in Dubai from Vestas everything is included: for example, Siemens, whomever, your switch gear, your gearbox, your generator, everything is interlinked and it behaves like one system. In PV that doesn’t happen. We need to bring suppliers into working together much earlier. So right now we have like 50 systems that are working independently, one from each other. So we need to bring this to the table. For wind, we’ve gone a step ahead and there’s a lot of predictive maintenance that is being already done. Machines diagnose themselves, know when something is going wrong when they need to stop when they need to restart. In PV we’re still figuring out what’s happening in our field because of the lack of sensoring, the lack of good quality data, the lack of knowing what’s going on. So definitely we need to bring all the experience and the come together and work together that was done in wind and bring it over on to PV.
Pritil Gunjan (Navigant Research): I’m coming back to Baywa, Virgil. I think you would probably have a couple of comments to make on how we can share best practices from wind to solar?
Virgil Cazacu (BayWa r.e.): Monitoring both technologies and doing the management, technical and commercial management, we have the luxury to exchange internally the experience. Because it’s a digital panel I would maybe stop to one example that’s again about buzzwords: A.I. and machine learning. They started a bit earlier in terms of developing self-services around the sensors from the turbines. I think what we can learn from there is that even the data which is collected from a turbine, which has hundreds of sensors, you want to do predictive maintenance on different components, a gearbox, IT systems and so on. And even if you have a lot of data coming from these sensors the results, which machine learning A.I. can bring, can be very skewed by the context. It’s always based on how you train the data sets by the experts from that wind industry. And now projecting to the PV side, where you have far less sensors and even the ones you have them on-site, most of the time they are decalibrated or sending the wrong data, it makes this even more complicated. All these future systems and algorithms which will give the success that we are looking for by standardization across components of the plants and the better integration between them will help us. That’s one of the lessons that we have learned.
Pritil Gunjan (Navigant Research): Thank you, Virgil. Johannes, any thoughts to share on how we can leverage wind digitization?
Johannes Burgard (Solytic): I’ve been sailing twice, but my experience in wind is very limited, to be honest. I think also the time is up.
Pritil Gunjan (Navigant Research): Can we have one last comment from Günter, if that’s okay?
Günter Maier (Alteso): On this topic or on any topic or on this wind topic you mean?
Pritil Gunjan (Navigant Research): No, on best practices or anything else, if you would like to add something.
Günter Maier (Alteso): Best practices I see in a way that – I repeat just what I think the gentleman from Wood Mackenzie first said – in order to move forward, it is important that all stakeholders are kind of having the interest to work together. That sounds always a bit boring, but I think here it is very much true because the value of data and insights exists only when it is used. It doesn’t matter when one stakeholder has data and insights, but other stakeholders have no access to it. Then nobody can benefit from progressing. We are trying to work together with other companies. For example, I’m very happy that we have a project with Above, which is a drone company, where we are trying to correlate the permanent analytics results with the information you get from drones. We believe this is a very good fit. I believe that sooner or later this is merging together anyway.
Pritil Gunjan (Navigant Research): Excellent. With that, I would like to thank our amazing panel today. Do we have any questions? Okay. So please feel free to reach out to them later. And thank you all for your patience.