Google DeepMind, in a nutshell, is a cutting-edge artificial intelligence research laboratory, a subsidiary of Google.
Google DeepMind, the pioneering AI research lab behind AlphaGo and other groundbreaking projects, might seem like an impenetrable realm accessible only to seasoned engineers and researchers.
They focus on advancing the field of general-purpose AI (AGI), also known as artificial general intelligence, which aims to create machines with cognitive abilities similar to humans.
What makes Google DeepMind special?
Here’s a breakdown of what makes DeepMind special:
Pioneering AI research: DeepMind has made significant breakthroughs in AI, particularly in areas like:
- Deep learning: Their algorithms have mastered complex games like Go and StarCraft, surpassed human performance in protein folding prediction (AlphaFold), and generated human-quality creative text formats (AlphaStar).
- Reinforcement learning: DeepMind uses this technique to train AI agents to solve problems and make decisions through trial and error.
- Neuroscience-inspired AI: They research how the brain works to develop biologically plausible AI models.
Real-world applications: DeepMind’s research isn’t just confined to the lab. It’s finding practical applications in:
- Healthcare: Tools like Streams use AI to predict and prevent acute kidney injury in patients.
- Energy efficiency: DeepMind helps Google data centers save energy by optimizing cooling systems.
- Scientific discovery: AlphaFold is revolutionizing protein research, potentially leading to new drug discoveries.
Ethical considerations: DeepMind also recognizes the ethical concerns surrounding AI and actively participates in discussions and research on safe and responsible AI development.
Google DeepMind Engagement opportunities:
Even if you’re not an AI expert, you can still engage with DeepMind:
- Explore their website and social media to learn about their research.
- Use Google products powered by DeepMind, like AlphaFold or personalized recommendations.
- Participate in citizen science projects like Perceiver IO.
Remember, Google DeepMind is at the forefront of AI research, pushing the boundaries of what’s possible. By understanding their work and engaging with it, you can stay informed about the future of artificial intelligence and its potential impact on our world.
But fear not, there are ways to engage with DeepMind’s work and even leverage its advancements in your own endeavors, regardless of your technical background.
Exploring the DeepMind Landscape:
- Dive into the Research: DeepMind’s website is a treasure trove of information. Visit deepmind.com to browse through their blog, research papers, and project pages. You’ll find fascinating articles on topics like protein folding, climate change prediction, and AI safety, often presented in digestible formats.
- Follow DeepMind on Social Media: Stay updated on their latest breakthroughs and events by following their Twitter (@DeepMindAI) and YouTube channel. They regularly share informative videos, podcasts, and talks by DeepMind researchers, offering insights into their work and its potential applications.
- Engage with the Community: DeepMind hosts forums and discussion groups where enthusiasts and experts exchange ideas and debate the implications of their research. Consider joining the DeepMind Forum or Reddit community to participate in stimulating conversations.
Beyond Exploration: Taking Action with Google DeepMind:
- Utilize DeepMind Technologies: Explore Google products and services powered by DeepMind’s AI, such as the AlphaFold protein structure prediction tool or the personalized recommendations in Google Play. These applications showcase the practical impact of DeepMind’s research.
- Contribute to Citizen Science Projects: DeepMind engages the public in contributing to research through projects like Perceiver IO, where anyone can help train an AI model to understand the world through videos. Participating offers a firsthand experience in AI development and contributes to scientific progress.
- Get Involved in Educational Initiatives: DeepMind offers educational resources like the DeepMind Quest AI coding game and workshops to introduce students of all ages to the principles of AI. Encourage kids and teens in your life to explore these resources and ignite their passion for AI.
Remember: DeepMind’s work is still under development, and its potential implications are vast and complex. Engage with a critical and informed perspective, participate in discussions around ethical considerations, and use your newfound knowledge to advocate for responsible AI development.
While DeepMind doesn’t have a direct user base, its impact extends to a vast number of people through its integration into various Google products and services.
Google DeepMind: Star in the Deep Learning Universe
Google DeepMind has carved a prominent niche in the ever-evolving landscape of deep learning. Their contributions extend far beyond the realm of cutting-edge research, leaving a lasting impact on various fields and inspiring future advancements. Here’s a peek into DeepMind’s significant role in deep learning:
Pioneering breakthroughs:
- Mastering complex games: DeepMind’s algorithms, like AlphaGo and AlphaStar, conquered the intricate worlds of Go and StarCraft, pushing the boundaries of AI’s strategic and decision-making capabilities.
- Revolutionizing protein research: AlphaFold, a deep learning tool for protein structure prediction, has transformed the field of biology, accelerating drug discovery and shedding light on protein functions.
- Generating human-quality creativity: AlphaStar’s ability to generate creative strategies in real-time gaming showcases the potential of deep learning for artistic expression and human-like problem-solving.
Advancing deep learning techniques:
- Refining reinforcement learning: DeepMind’s research has significantly contributed to the field of reinforcement learning, enabling AI agents to learn through trial and error, leading to impressive feats in game playing and robot control.
- Developing scalable architectures: They’ve designed efficient and adaptable neural network architectures like Perceiver IO, capable of handling diverse data modalities and scaling effectively to large and complex tasks.
- Integrating neuroscience insights: DeepMind’s research often draws inspiration from how the human brain works, resulting in more efficient and biologically plausible AI models.
Impact beyond research:
- Real-world applications: DeepMind’s AI powers tools like Streams for predicting and preventing acute kidney injury, demonstrating the potential for improved healthcare using deep learning.
- Optimizing energy efficiency: Their AI aids Google data centers in reducing energy consumption, highlighting the environmental benefits of deep learning-based solutions.
- Democratizing AI: Projects like Perceiver IO strive to make deep learning tools more accessible and adaptable, paving the way for broader adoption and innovation.
Challenges and considerations:
While DeepMind’s achievements are noteworthy, it’s crucial to acknowledge potential challenges:
- Ethical concerns: Issues like bias, transparency, and job displacement require careful consideration and responsible development of deep learning applications.
- Accessibility and equity: Ensuring equitable access to the benefits of deep learning and mitigating potential inequalities is crucial.
- Long-term implications: The potential far-reaching consequences of general-purpose AI necessitate ongoing research and ethical discussions.
Looking ahead:
DeepMind’s pioneering work in deep learning continues to inspire and shape the future of AI. Their dedication to tackling complex challenges, advancing the field responsibly, and exploring real-world applications positions them as a leading force in shaping a more intelligent and beneficial future for all.
The story of DeepMind in deep learning is far from over. As research progresses and technological advancements unfold, one thing is certain: DeepMind will remain a prominent force, illuminating the path towards a future where AI enhances and enriches our lives in innovative and transformative ways.
Google DeepMind Audience
Here’s a breakdown of how DeepMind’s technologies reach a wide audience:
- AlphaFold: This groundbreaking protein structure prediction tool has been used by over 100,000 researchers worldwide, accelerating drug discovery and advancing our understanding of biology.
- Google Play: DeepMind’s AI powers personalized recommendations in the Google Play Store, reaching over 2.5 billion active Android devices.
- Google Assistant: DeepMind’s research has contributed to advancements in natural language processing, making the Google Assistant more responsive and helpful for millions of users.
- Google Data Centers: DeepMind’s AI has helped to reduce energy consumption in Google’s data centers by up to 40%, indirectly benefiting billions of Google users by reducing the environmental impact of their online activities.
Beyond these direct applications, DeepMind’s research has also influenced other AI projects and advancements across various industries, further extending its reach.
The benefits of Google DeepMind
The benefits of Google DeepMind are far-reaching and multifaceted, impacting both individuals and society as a whole. Here’s a glimpse into some key areas:
Scientific & Technological Advancement:
- Accelerated scientific breakthroughs: DeepMind’s AI tools like AlphaFold are revolutionizing protein research, leading to faster drug discovery and development.
- Enhanced problem-solving capabilities: By tackling complex challenges in areas like materials science and climate change prediction, DeepMind is paving the way for innovative solutions.
- Automation and optimization: From optimizing energy efficiency in data centers to streamlining logistics networks, DeepMind’s AI is driving automation and increasing efficiency across various sectors.
Societal & Individual Impact:
- Healthcare improvements: AI tools like Streams help predict and prevent acute kidney injury, potentially saving lives. DeepMind’s research also contributes to personalized medicine and early disease detection.
- Enhanced quality of life: AI assistants powered by DeepMind can assist with daily tasks, provide reminders, and even offer companionship, particularly for vulnerable populations.
- Educational opportunities: DeepMind’s educational initiatives, like DeepMind Quest, spark interest in AI among young people and provide valuable learning experiences.
Economic Growth & Productivity:
- Increased productivity and automation: DeepMind’s AI can automate repetitive tasks, freeing up human workers for more strategic roles and boosting overall productivity.
- New job opportunities: The development and implementation of AI will create new jobs in areas like AI development, maintenance, and ethical oversight.
- Economic expansion: By driving innovation and efficiency across various sectors, DeepMind can contribute to economic growth and development.
However, it’s important to acknowledge that DeepMind’s advancements also come with potential challenges and risks:
- Ethical considerations: Issues like bias, transparency, and job displacement need careful consideration and responsible development of AI.
- Accessibility and equity: Ensuring equitable access to the benefits of AI and mitigating potential inequalities is crucial.
- Long-term implications: The potential far-reaching consequences of general-purpose AI necessitate ongoing research and ethical discussions.
Google DeepMind’s Perceiver IO is a groundbreaking project that pushes the boundaries of what’s possible with neural networks. It’s a powerful and flexible architecture capable of handling a wide array of tasks with diverse inputs and outputs.
Here’s a breakdown of Perceiver IO and its key features:
What it does:
- Handles diverse data: Perceiver IO can process different kinds of data like text, images, audio, video, and point clouds, making it truly versatile.
- Scalability and efficiency: Unlike traditional Transformer models, Perceiver IO doesn’t suffer from quadratic complexity as input size increases. This means it can handle large and complex inputs efficiently.
- Structured inputs and outputs: Perceiver IO excels at tasks with structured inputs and outputs, such as predicting optical flow between images or generating melodies from musical notes.
- General-purpose architecture: It’s not limited to specific applications and can be adapted to various tasks, making it a valuable tool for researchers and developers.
How it works:
- Cross-attention and latent space: Perceiver IO uses cross-attention to project high-dimensional input data onto a lower-dimensional latent space. This compressed representation is then processed by a Transformer module, making it more efficient.
- Flexible output: The latent space is then decoded into the desired output format, enabling Perceiver IO to generate outputs of various shapes and sizes.
Impact and Applications:
- Real-world potential: Perceiver IO has shown promising results in tasks like image inpainting, protein structure prediction, and robotics control.
- Democratizing AI: By being adaptable and efficient, Perceiver IO can pave the way for more accessible and versatile AI tools.
- Unlocking new possibilities: Its wide range of applications makes it a valuable tool for various fields, from scientific research to creative endeavors.
20 Fascinating Google DeepMind Projects
1. AlphaGo & AlphaStar: These AI programs conquered the challenging worlds of Go and StarCraft, respectively, showcasing DeepMind’s prowess in game playing and strategic decision-making.
2. AlphaFold: This groundbreaking tool predicts the 3D structure of proteins with remarkable accuracy, accelerating drug discovery and unlocking new insights into protein function
3. Gato: This multi-modal AI agent demonstrates impressive versatility, excelling in different tasks like playing Atari games, generating text, and controlling robotic limbs.
4. AlphaZero: This self-play learning algorithm mastered chess, Go, and Shogi from scratch, highlighting DeepMind’s advancements in reinforcement learning and game playing.
5. MuZero: Building upon AlphaZero, MuZero can learn to play games and solve other tasks without any prior knowledge or human intervention, showcasing its general-purpose learning capabilities.
6. Streams: This AI-powered tool helps predict and prevent acute kidney injury in patients, demonstrating DeepMind’s potential to revolutionize healthcare with intelligent algorithms.
7. Perceiver IO: This flexible neural network architecture handles diverse data modalities and scales efficiently, making it a powerful tool for various tasks like image inpainting and protein structure prediction.
8. Deep Green: This project aims to develop AI for sustainable energy management, optimizing energy consumption in data centers and contributing to environmental goals.
9. Robotics & Control: DeepMind’s research in robotics control equips robots with intelligent decision-making and movement capabilities, paving the way for more flexible and autonomous robots.
10. Language Understanding: DeepMind’s advances in natural language processing enable AI to understand and generate human-like language, with applications in chatbots, machine translation, and text summarization
11. Multimodal Learning: By integrating information from different modalities like vision, sound, and touch, DeepMind’s AI gains a richer understanding of the world, leading to more robust and adaptable algorithms.
12. Unsupervised Learning: DeepMind explores unsupervised learning techniques where AI learns from unlabeled data, unlocking the potential to analyze massive datasets and discover hidden patterns.
13. Fairness & Ethics: Recognizing the ethical considerations of AI development, DeepMind actively researches and promotes responsible AI practices to mitigate bias and ensure fair and beneficial applications.
14. Explainable AI: DeepMind strives to develop AI models that are interpretable and understandable, allowing humans to understand the reasoning behind their decisions and build trust in AI systems.
15. Safety & Security: Ensuring the safety and security of AI systems is paramount, and DeepMind actively researches methods to identify and mitigate potential risks associated with advanced AI.
16. Democratizing AI: DeepMind initiatives like DeepMind Quest aim to make AI more accessible and engaging for everyone, fostering interest and understanding of this transformative technology.
17. OpenAI Collaboration: DeepMind’s recent merger with Google AI’s Brain division signifies a commitment to collaboration and accelerated progress in the field of AI, potentially leading to breakthrough advancements.
18. Future of Work: As AI continues to evolve, DeepMind actively explores the potential impact on the workforce and works to ensure a future where humans and AI co-exist and collaborate effectively.
19. Long-Term AI Safety: DeepMind recognizes the importance of considering the long-term implications of advanced AI and actively participates in discussions and research on safe and responsible.
Beyond the List:
This list of 19 projects scratches the surface of DeepMind’s diverse and impactful work. They constantly embark on new ventures, tackling complex challenges and unveiling the immense potential of AI across various fields. Whether it’s revolutionizing scientific research, optimizing energy consumption, or shaping the future of work, DeepMind’s contributions hold the promise of a better and more intelligent future for all.
Conclusion The Deepmind: How to Engage with Google’s Cutting-Edge AI
Google DeepMind offers immense potential for progress and positive impact across various aspects of our lives.
DeepMind doesn’t disclose specific user numbers for its research projects or internal tools. However, the widespread adoption of its technologies within Google products and its influence on the broader AI landscape demonstrate its significant impact on a global scale.
By stepping into the world of DeepMind, you can gain valuable insights into cutting-edge AI research, contribute to scientific progress, and even leverage its advancements in your own life.
https://www.exaputra.com/2023/12/the-deepmind-how-to-engage-with-googles.html
Renewable Energy
Pardalote Studies Australian Blade Erosion and Heat Fatigue
Pardalote Studies Australian Blade Erosion and Heat Fatigue
Rosemary Barnes, CEO and founder of Pardalote Consulting, joins to discuss their new grant-funded study of blade erosion and heat fatigue in Australia.
Sign up now for Uptime Tech News, our weekly newsletter on all things wind technology. This episode is sponsored by Weather Guard Lightning Tech. Learn more about Weather Guard’s StrikeTape Wind Turbine LPS retrofit. Follow the show on YouTube, Linkedin and visit Weather Guard on the web. And subscribe to Rosemary’s “Engineering with Rosie” YouTube channel here. Have a question we can answer on the show? Email us!
Welcome to Uptime Spotlight, shining light on wind energy’s brightest innovators. This is the progress powering tomorrow
Allen Hall 2025: Well, Rosemary, welcome back to the show.
Rosemary Barnes: Thanks, Allen. Great to be here. For, it’s been a while since we did one of these one-on-one episodes, like a, yeah, a proper, proper guest.
Allen Hall 2025: Well, this is kind of a celebratory episode because your company, Pardalote Consulting, has been awarded, uh, some funding from the Australian Capital Territory’s government for the Energy Innovation Fund.
Rosemary Barnes: It’s a really good program that the ACT government has to try and get energy innovation In the state. It’s not a state actually, it’s technically a territory. Little more than just Canberra, the city. Uh, but there are actually quite a few, like, really interesting energy-related companies here, partly ’cause of the, the fund I think helps, but also just tracing back like, [00:01:00] uh, y- you know, in the 20-teens, Australia had a really conservative government that hated renewable energy, and the ACT government had a commitment at that time to 100%, um, 100% renewable electricity for the, the government.
And that was one of the only programs that was resulting in a lot of, um, you know, clean energy projects being built, and one of the conditions that they put on that, uh, for people that would win PPAs with the ACT was that you had to have your headquarters in Canberra. So we’ve actually got quite a few, quite a few really cool, innovative companies out of here.
Um, like Neoen’s headquarters here. Windlab, uh, yeah, was, was founded here and still has a lot of people here. Pardalote obviously, and you know, a few other companies as well. So despite it being a small city of like, I don’t know, maybe it’s up to 400,000 or something people by now, um, yeah, there is actually quite a lot going on here for energy.
Allen Hall 2025: And the Energy Innovation Fund is funded by the wind and solar operators in the area, and your particular [00:02:00] effort has really global consequences. You’re focusing on two areas involving how wind turbines survive Australia, but more, uh, of relevance is to just really tough conditions which exist not just in Australia but around the world.
What two areas are you going to focus on?
Rosemary Barnes: Yeah. So the two focus areas are leading edge erosion and high temperature fatigue, which we can probably get into the definitions of those in a minute. But basically my, um– what led me to wanna have a project like this was that when I moved back to Australia in 2021, I– and I started working in O&M, uh, I noticed that the wind turbines that I would look at, the blades that I would look at here behaved really differently to the ones that I worked with overseas.
You know, es- especially with leading edge erosion, like often I would be doing a condition assessment of a, you know, a new wind farm. Um, might only have been operating for, you know, two years. That’s a pretty common time for people to get in and do a condition assessment [00:03:00] because their warranty period is about to end and they wanna, you know, make sure that everything is okay.
Um, and I would just notice that often, like 90, 100% of blades would already have bad erosion after just a couple of years, which is super-duper fast. And then there are some tools available to check, um, like what kind of erosion are you likely to experience on your site. Like is it a higher severity erosion site or a, a low severity one?
Um, and you basically, you know, the status quo globally is to just look at the annual rainfall, um, and the tip speed. And if you’ve got, you know, high for both of those, that’s a bad erosion site. And if you’ve got low for both of those, it’s a, a low erosion site. But when I plotted out the wind farms that I knew had really bad erosion problems onto, you know, a chart with those two axes, I just saw a random distribution of dots.
You know? Like, this was not– uh, this had no predictive value for Australian wind farms. And so that led me to believe that, okay, um, you know, things are a bit [00:04:00] different here. Makes sense, you know, most of the knowledge that we have about how wind turbines operate, it’s been developed and validated mostly in Northern Europe.
You know? Like it’s, it’s Denmark and the surrounding countries that had, like, the bulk of the early wind energy. First few decades of knowledge were, you know, were mostly there. Of course, there were some other, um, places that had wind turbines, but, you know, most of the The OEMs have been operating for decades, came from Denmark.
And I know when I lived in Denmark, the rain there is very different to the rain in Australia. So in Denmark, it’s basically always raining, right? Like, it’s just… Like, even if it’s not raining, you’re still gonna get wet when you go outside ’cause it’s just, like, the air has this just amazing ability to just hold onto moisture.
Um, but it’s very, very gentle. But, you know, over an entire year of most days having gentle rain, that adds up to a lot. Whereas in Australia, and especially if you go, like, north to Queensland, it rarely rains. It’s mostly just dry, and when it [00:05:00] does rain, it’s like a tap turns on, and I, I swear you will get bruised from the rain droplets hitting your skin.
You know, they just have so much energy in them. So I think that that i- you know, when you look at just the overall rainfall, you really hide something important about how erosion, um, can progress. Then, um, there’s other places in Australia that have very different characteristics. Again, they don’t have that kind of really intense rain but, you know, some of those sites are also having really bad erosion.
And so it just occurred to me, I did a lot of research, you know, into what’s going on and, you know, the academics are studying erosion a whole lot, and they’ve got, you know, a lot of standardized tests and, you know, products are developed according to these standardized tests. But the standardized tests don’t actually resemble reality, and especially they don’t resemble reality in Australia.
And so my client started asking me, “Okay, you know, the products that we have are, are terrible. We have to replace them every couple of years. It’s, um, causing big problems with also [00:06:00] the amount of energy that you’re losing.” One of the types of, um, leading-edge erosion or leading-edge problems that we have in Australia is that the, the coatings tend to peel off and make these, like, big flakes which will just massively disrupt the airflow, can cause y- you know, at least a few percent AEP loss, and maybe up to five.
And even worse than the AEP loss is the revenue loss because it affects it most at, you know, lower wind speeds. Um, you get a bigger hit than at rated wind speeds. So there’s a variety of problems going on with leading edges in Australia, which mean that I, I basically… My clients would ask, “What product should we put on to prevent having to, you know, constantly replace this?”
‘Cause it costs, like- you know, 30, $40,000 per turbine to replace the protection, not to mention, you know, one or two days of downtime. It’s expensive, and I basically, I didn’t have a good answer for them. What, what product should they put on? I don’t know. No, we, we don’t know. One, we don’t know what the [00:07:00] specific, um, characteristics are that are…
what the specific local environment, local conditions are that are accelerating leading-edge erosion, one. And two, all of the products tend to be tested around this, you know, there’s this protocol that academics have come up with, and they’ve kind of like assumed that this is representative of how things behave in the field, and it’s– I don’t think it’s particularly true anyway, but it’s especially not true in Australia.
There are a few companies that are testing to different standards. Um, definitely applaud them. But without knowing wha- what are the conditions truly like in Australia, uh, it’s really hard to advise, like, what kind of tests should you be demanding from a product you’re considering to be sure that you’re gonna put it on and not gonna be replacing it again in two years.
Allen Hall 2025: Because that’s really the trouble in Australia is when you get offered products They have been tested generally in somewhere in Europe and maybe in the United States, and then when they go to [00:08:00] Australia, it’s really unknown as to how those products will do, which is a huge risk for the Australian wind market as to what to choose, how to choose, is it– what’s real in terms of test data.
So now you’re gonna go out and do what? Are you gonna put sensors out by the wind farms? Are you gonna try to do more of a statistical summary of the actual environment around wind farms using existing data? What’s the approach here?
Rosemary Barnes: It’s all of the above, but the part that is supported by the grant is that we’re gonna have enough money to be able to buy some scientific-grade sensors and put them on, um, a sample of Australian wind farms.
So we’re gonna be looking at a lot more characteristics about the rain than simply is it raining now, you know, how many millimeters per hour. We’re also gonna be investigating, you know, every kind of characteristic of, of that, um, of that rain, um, including, yeah, like the, the energy that’s in it, for example.
A, a bunch of stuff. I won’t get into every single [00:09:00] parameter. Um, and you know, other things as well, like measuring UV, solar radiation, um, particles, because, you know, in Australia we have a lot of dirt roads, which I know is very common in wind farms around the world, but Australian dirt roa- roads are always dry and dusty, like 99% of the time, so that’s one of the things that y- you know, maybe that’s causing a difference.
Um, so basically putting sensors all over a bunch of wind turbines and then monitoring the erosion, um, a combination of some real-time monitoring and also looking at inspection, um, drone inspection images annually. We also have a- an option where we’ll just be using SCADA data and inspection images, so that’s like a lower cost version where we can combine that with the findings from the scientific-grade instrumented turbines to build up a picture of what types of conditions lead to accelerated erosion.[00:10:00]
Allen Hall 2025: So the SCADA data will, will have some information inside of it, you think, that, uh, will correlate to the weather outside?
Rosemary Barnes: It has some Additionally, we can look up, um, you know, just the weather data, like how many millimeters fell during which 15-minute interval throughout the day, what was the temperature.
SCADA will tell us also what the temperature was, um, what the speed of the turbine was, so you can calculate the tip speed, ’cause that’s an important thing. Um, yeah, so it’s, it’s two, it’s two tiers of data collection. The scientific grade sensors, as you can imagine, are, are really expensive and y- you know, the, the grant project has contributed a, a lot of funding, um, but it’s not enough to put those, yeah, put a little mini lab on top of every turbine across Australia, obviously.
So that we’re using s- doing selectively, and then we can increase the number of wind farms that are included in the study by just doing this, um, cheaper version of the SCADA [00:11:00] plus, uh, weather data that’s available.
Allen Hall 2025: So what are some of the risks on the temperature side for all the high-temperature regions of Australia that have wind turbines?
Clearly it’s generally warmer in Australia than it is in, in Scandinavia and Northern Europe. What kind of temperatures are we talking about on the ground?
Rosemary Barnes: Uh, well, temperatures here can get pretty close to 50 degrees. Um, and if you’ve ever been inside a wind turbine blade on a, even a mildly hot day, you’ll know that the temperature inside a wind turbine, and especially inside the blade, is much hotter than what it is, uh, what the ambient temperature is.
So this project is one– I’ve actually been talking about this project for, yeah, like over 10 years now. Ever since I started, I moved to Denmark, started working for a wind turbine manufacturer, I had done– I had just finished doing my PhD on composite materials, structural design, and analysis. So, um, yeah, very, very familiar with, [00:12:00] you know, how composite materials work and, in particular, the effect that temperature has on them.
I mean, like most materials, when composites get warmer, they get softer, and that is really important for a w- a wind turbine blade. You know, if it gets, um, less stiff, then you’re gonna get a lot more strain, and that is going to affect your fatigue behavior. Y- you know, fatigue is just the application of a little bit of, a small amount of strain.
It’s not gonna cause damage, but when you apply it millions, tens of millions of times, like you do in a, o- over a wind turbine’s operate, um, operating lifetime, then that builds up. And, you know, wind turbine blades are a very fatigue-driven design. Um, it’s one of the most important things to consider when you’re designing a wind turbine blade.
And so when I got to Denmark and I learned how materials are qualified and how the qualification is treated in the certification process, I just realized it’s not particularly conservative, and also that some of the assumptions that are made that [00:13:00] wo- again, they worked really well in more moderate climates where wind turbines have had most of their developmental history.
You know, it’s not such a big deal there if you test at room temperature. Your wind turbine blade is spending most of its operating lifetime at room temperature or below. It’s, it’s rarely, you know, above 30 degrees in Denmark and most of Northern Europe and, you know, also a lot of, um, a lot of America, not, not all of it But, um, in Australia it has just extended periods above that temperature and even exceeding the temperature where, you know, wind turbines have an operating limit and after that they will shut down.
But the operating limits are based on ambient temperature. It’s not based on what’s the temperature in the laminate, which is what really matters for blade lifetime. So anyway, I’ve been obsessed, like honestly obsessed about this issue for 10 years. Talked about it with anybody who would listen . But then when I started working in O&M in [00:14:00] Australia and I started seeing some wind farms with an abnormal number of cracks early…
again, early in their lifetime, you know, I think one of the wind farms I was looking at was maybe three years old or four at the time. I think it was three actually, and had a lot of cracks, and I looked at a few years in a row and it was more and more cracks every year and I’m like, “Oof, this really looks like end of life fatigue behavior.”
A- actually it’s not, y- you know, there’s this concept of a bathtub curve where, um, when you’re looking at failures in components, in, in anything, not just in, um, wind turbine blades, but you know, like you’d start– it’s called a bathtub because, you know, when it starts operating, you’ll get quite a lot of failures.
Anything big, any manufacturing defects or anything are gonna cause failures quite fast, and that kind of drops off over time as all of those, uh, get addressed. And then you have, you know, the bulk of your operating life, it’s like pretty low level, pretty, pretty constant for a long time and then as you get towards the end of the [00:15:00] life, you start to see failure rates rise up again.
That’s your fatigue failures, your end of life fatigue failures. And so when I saw the same types of cracks more and more each year, I’m like, “This looks like, you know, the foot end of the bathtub, not the head end.” And, uh, it made me worried and I’ve now seen that across a few wind farms in Australia at, um, hotter places.
There’s a few blade types that are more prone to it than others, but at this point it’s still a suspicion that that’s what’s going on. I mean, a suspicion backed by a lot of, a lot of theory and knowledge of how the certification process works. But this project now we’ve got some funding to actually go put some sensors onto wind turbines, actually learn what the temperatures are in the blades throughout the whole laminate, um, not just the, you know, on the outside surface or not just the ambient temperature, but actually, you know, develop a temperature gradient across the whole, um, the whole laminate in the blade shell.
Um, and [00:16:00] then we’re going to be doing a bunch of modeling basically to look at what is the effect of these different temperatures that blades are really seeing and how much would we expect to… that to decrease a lifetime. And then we should also be able to say, you know, if you have this issue in your wind farm, you might be able to change your operation a little bit and extend your lifetime a lot.
Because this one, it’s real– like, in contrast to leading edge erosion, leading edge erosion is just, it’s, you know, every wind turbine has it to a certain extent, and it, it’s always there, but it’s a relatively minor cost to fix it. You know, like it sounds like a lot, like 30, $40,000 per wind turbine, but, um, you know, compared to if you’ve got to replace every blade across your fleet because they’re all, you know, at the end of their life after five years, you know, that’s obviously shocking.
And, you know, that’s a bad example, but even in a y- you know, like a less extreme example, maybe [00:17:00] after 15 years you have to do a, you know, a f- a fleet-wide campaign to strengthen blades or something. It’s, you know, m- many millions of dollars for that, and so it c- could make sense to be able to learn, okay, what, what hours of operation should we be avoiding?
Additionally, because when it’s super-duper hot in Australia, usually you’ve got heaps of solar power and the electricity price is not that high. So I, I think that there– and I don’t, obviously, before we’ve done the project, I don’t know what the threshold is. But in both cases, we will be aiming to improve the knowledge of how you can operate to avoid these periods of accelerated damage.
Allen Hall 2025: Do you think you’re seeing more fatigue-like damage due to the blades operating when it’s hot or not operating when it’s hot, with maybe less airflow around the blade and maybe less cooling going on is just a temperature soak At rest? [00:18:00]
Rosemary Barnes: Yeah. It’s interesting because the temperature is higher if it’s not rotating, um, because you get a whole lot of, um, convective heat, heat transfer when the turbine is operating.
So your temperatures are not gonna get as hot when operating as when they’re standing still. However, if it’s standing still, they’re only very lightly loaded. Like, yes, they’re gonna get, um, blown by, by gusts and, um, have a little bit of bending, but it’s, it’s very, very small compared to, uh, if it is y- you know, operational loads.
Uh, assuming that you’re not in the middle of a s- a storm. But yeah, a storm probably doesn’t come with 50 degrees temperatures.
Allen Hall 2025: And what part of the blade is susceptible to these higher temperatures? Is it the resin? Is it the fiberglass or carbon fiber? Or is it the, the glue, the bond joints? What part are you focused on?
Rosemary Barnes: The resin is the main part that I’m focused on. It gl- it could be an issue for glue too, actually. I haven’t even looked into what the, um, yeah, temperature assumptions are with, with glue, with [00:19:00] bond lines. But the failures that I’m seeing in the field are not, are not bond line issues. It’s, it’s, um, a laminate problem.
Allen Hall 2025: What about balsa and foam inside of the blade? Are they affected by the temperatures or are they pretty temperature stable?
Rosemary Barnes: I don’t think they’re affected at these kinds of temperatures, no. They, they don’t really do much actually. The, the core materials, like it, it is very important that they’re, that they’re there, but their job is really to keep the fiberglass separated from its- itself to make it stiffer.
So, um, yeah, that’s, that’s unlikely to be a, a major source of problems.
Allen Hall 2025: So this study is gonna work over about three years, and you have a number of wind farms that are participating. Are you looking for more wind farms to participate in Australia?
Rosemary Barnes: Yeah. Yeah, definitely. I mean, we can, um, have as many as, as people want to join.
We’ve got quite a good selection so far. Definitely can always welcome more. A, a bit limited in how many can get the really, um, good sensor [00:20:00]package, because the grant funding is a, you know, a certain amount, and that’s paying the bulk of those sensors. So, um, those spots are limited. So if anybody wants to really zone in on what is specifically causing erosion on their site, you know, if you know that you have got leading edge protection that is not good enough and you have to replace it soon, but you don’t know what to replace it with, then, you know, that would be the kind of wind farm that might want to consider, yeah, joining this and, um, you know, getting these sensors on their, um…
We’re putting them on top of the nacelles, most of them. Um, yeah, so that would be a good match then. Um, and then, yeah, for the ones that are doing the SCADA data and, um, weather data- There’s not such a, a hard limit on how many we can have join like that. So yeah, we can have more, more like that.
Allen Hall 2025: In the temperature fatigue effort, i- is that still looking for participants or are there particular wind turbine types or manufacturers that you’re [00:21:00] looking for to participate?
Rosemary Barnes: Yeah, I think, um, I, I mean yes, we can have more of those. That’s a simpler, a, a simpler issue as well. The sensors are not so expensive and, um, it’s, yeah, it’s a, it’s a simpler project to join that one. We only need, you know, a couple of turbines per site, so it won’t be such a, uh, an involved process to get everything up on into the turbines.
And in terms of who might like to join that, I would say anybody that is in a really hot area where, you know, where they see a lot of days over 30 degrees, and if they see any days, you know, getting into the high 40s, then I would say that that’s worthwhile. Or even I have seen this issue in some milder sites, um, yeah, depending on the, on the blade type as well.
It is more common with polyester resins. They have a, a lower op- uh, maximum operating temperature than epoxy resins. But then also just anybody that has noticed just, hey, [00:22:00] we’ve got a lot of cracks, and it seems like we’re getting more and more cracks every year, which to be honest, can be hard to keep track of if you’re…
If you’ve got a full service agreement, uh, you know, an OEM managing your wind farm The early signs of this are gonna be category one and category two cracks. They’re not in exactly the same location. It’s, you know, it’s a tricky one. Normally, if you’re looking at a serial issue, then you’re going to have, uh, well, you know, your ideal pattern for a serial issue is the exact same thing happening over and over again.
And so it is harder to pull this out. It also really would be very rare for it to be happening in the first two years or three years, whatever your serial defect liability period is. So it’s quite hard. But, um, another group of wind farms that might like to consider it is if you know that in, you know, a certain number of years you have to renegotiate your service agreement or, you know, it ends and you might have to take over yourself, then this’ll be a really good way for you to [00:23:00] understand, you know, have I got a ticking time bomb here?
Um, because it’s not something that you’re gonna be aware of if you haven’t been, you know, doing some really, really in-depth shadow, shadow monitoring of your blades, you know, running your own inspections and looking at every single damage, not just category three, four, five, but lower ones. So yeah, I mean, there’s a, a wide variety of people that, that could be interested in joining.
Allen Hall 2025: Are you expecting a number of manufacturers that make leading-edge protection or involved in resin creation, some– there’s a number of resin companies and a variety of resins that are used globally, sort of interchangeably at times. Are you expecting some of those companies to participate in this effort just to learn about the Australian environment?
Rosemary Barnes: I think it would be a good opportunity to test out some products and see how they behave in the Australian context. I think that that would be a really good selling point, but I, I have to say that most of the companies doing that sort of thing that wanna enter Australia, they don’t [00:24:00] really consider…
Like, from the perspective of wind farm owners in Australia, if you can’t show us wind farms in Australia where this has worked and, you know, show us a before or after, you know, the old LEP lasted Two years and our LEP is going on four years now with no damage. It, you know, unless you’ve got a before and after like that, you can tell us however many turbines that you’ve got installed around the world, but, um, we don’t consider it validated, y- you know?
It’s not validated for Australian conditions yet. And I do have this same discussion over and over again with, you know, not just leading edge protection, but all kinds of, um, you know, manufacturers of whatever doodads that you put on to improve a, a wind turbine. It’s so different to Australia. Things break so fast.
And I’m talking everything, you know, like vortex generators fall off and, um, yeah, like, uh, you know, bits of lightning protection systems fall off, seals just [00:25:00] crumble and disintegrate. Um, and it, you know, we’re very wary of, of new products. So I, I do– I mean, I’m thinking of it more from my client’s point of view than from the product manufacturer’s point of view.
But one thing that I wanna get out of this pro- project is to be able to answer one of the most common questions that I get is, which is, what leading edge protection should I be putting on my turbine? And for now, I don’t know. I, I know a range of products that don’t work in Australia, and not much more than that.
So, um, yeah. And it’s also, you know, Australia’s a very varied place with lots of different kinds of climate too. So it’s not gonna be like, you know, the product that works in Queensland is the same one that’s gonna work in Tasmania, which is the same one that’s gonna work in Western Australia. You know, um, so it, this project is gonna really pull out what are the site specific issues you’ve got at your site and what kinds of, um, you know, tests would we need to see a product um, perform in order to know that this [00:26:00] is gonna last on your site.
Allen Hall 2025: W- what is the outcome of this project or these two projects? Are they gonna be reports or, uh, a, a continual monitoring system that’s designed for the Australian environment? How do you see this going?
Rosemary Barnes: Yeah, so one part of it is, um, developing a way to identify periods of accelerated damage and to know not to operate during that time.
So we call it protective operation. Uh, so that would, uh, help you if, yeah, you’re trying to extend the life of something or increase the amount of time before you have to repair, then y- you know, that would be useful to have that knowledge. And it will be as simple as just an alert saying, “Hey, accelerated damage conditions.
Consider, you know, if you wanna keep on operating.” And, you know, if the price of electricity is super high at that time, they may want to push through, and if it’s low, they probably won’t want to. So that’s one thing. Um, especially, you know, as wind turbines get to their, near the end of their life. I’ve got some clients whose wind farms only have, you know, [00:27:00] maybe five years operation left.
They just simply don’t wanna repair their leading edge protection again. They just, they, they don’t wanna do that. So they would be happy to, you know, reduce operation a bit and have their turbine limp through to the end of the period. Y- you know, you want everything to wear out at once. You don’t want brand-new leading edge protection on a turbine that’s going to come down in a couple of years.
Um, so, you know, that’s, that’s one part of it. And then the other thing is, you know, turbines earlier in their lifetime, how can we optimize the maintenance schedule with leading edge erosion? Um, so, you know, like it’s a lot cheaper to, uh, replace the LEP if you get– catch it early, but then you don’t wanna be catching it too early and replacing it, you know, constantly when you, you don’t need to.
So, um, yeah, it, this, having this knowledge will enable a site-by-site operations and maintenance strategy with respect to leading edge protection. We also have some sites who are having trouble. They’ve got a full service agreement, and the OEM is [00:28:00] responsible for, um, doing the leading edge erosion repairs and protection replacement, but the owner is on the hook for paying for it.
At the other end, we’ve got people with full service agreements where technically the, um, manufacturer is supposed to be doing the leading edge protection and paying for it, but they argue about what, when does it need to be done. Because, you know, um, the operator might think if there’s no structural risk, then we don’t need to be replacing it.
And in the meantime, you’ve got turbines spinning around for years and years and years with, you know, these huge flakes of leading edge protection s- you know, causing the flow at the tip of the turbine to, to detach and to stall, and horrible aerodynamics, huge losses in power generation and revenue. And they’re having a big fight about, you know, is this necessary to do or not?
And then, you know, they’re just gonna put the exact same product on again ’cause the [00:29:00] OEMs are re- all really, really wedded to their own particular brand. It’s like, “Well, last time we had this product and it was factory applied, it lasted one year before it s- it was worse than, you know, if it wasn’t there at all.
Uh, we don’t really want you to put that one on again.” And so, you know, having the information that they need to be able to, you know, really bring data to these discussions and, you know, makes a, yeah, data not drama. That’s a, a good approach I think, um, for any kind of negotiation and especially in the case of leading edge erosion.
And then for the high temperature fatigue part of the problem, aside from, you know, just wanting to know are your blades aging, should you be looking at remediation action or changing the operation, the other really big key thing is, uh, you might need to have a fight with y- your OEM about if this turbine has been designed and operated correctly.
And so then having the data from this, um, project is going to give you the information that you need to come into that [00:30:00] argument with, again, the data not the drama. Um, and to, you know, in- increase your chances of succeeding in that kind of really tricky negotiation.
Allen Hall 2025: So if you’re an OEM or a manufacturer of equipment, an ISP, an operator, pretty much all aspects of wind operations, you probably ought to be getting a hold of Pardalote Consulting and Rosemary to talk about the opportunity to participate in this study.
How do people get ahold of you to, to do that?
Rosemary Barnes: People can go to our website, pardaloteconsulting.com, and get in touch via the contact form there, or you can, uh, look me up on LinkedIn, Rosemary Barnes. That’s probably the easiest, fastest way to get ahold of me personally.
Allen Hall 2025: Well, Rosemary, congratulations on the Energy Innovation Fund Awards and the new three-year effort.
If you are interested in participating with Pardalote Consulting and working with Rosemary and her team [00:31:00] in Australia, reach out to her on LinkedIn and get that process started, because this report and the data from all this analysis that’ll happen over the next couple of years will be important to the wind industry.
So you need to spend some time and get ahold of Rosemary and get this process started now. So Rosemary, congratulations. Uh, thanks for being back on the podcast, and looking forward to, uh, the next couple of years. It sh- should be exciting.
Rosemary Barnes: Thanks so much, Allen.
Renewable Energy
Artificial Stupidity?
We all understand that there are ultra-conservatives living all around us, but does anyone truly believe that our schoolteachers are ruining our society by teaching children the truth about U.S. and world history? Science? Current events?
Slavery and Jim Crow laws were bad. Fascism is bad. Our scientists are telling us that CO2 emissions are causing world temperatures to rise, destroying our planet’s capacity to support life.
Whom do these concepts upset?
Renewable Energy
No Such Thing as a “Dumb Question”
There is nothing dumb about the question posed at left. Democracies fail, falling into “banana republics” constantly. The rate at which democracies become tyrannies is so great that some of them never make the news. Can you tell me anything about the governments of Eritrea or Chad?
What makes the situation in the United States is, yes, that it’s happening here in the United States, the very last place anyone would have suspected it.
You might have thought that Americans wouldn’t have voted for their nation to become Russia or North Korea.
You would have been wrong.
-
Greenhouse Gases12 months ago
Guest post: Why China is still building new coal – and when it might stop
-
Climate Change12 months ago
Guest post: Why China is still building new coal – and when it might stop
-
Greenhouse Gases2 years ago嘉宾来稿:满足中国增长的用电需求 光伏加储能“比新建煤电更实惠”
-
Climate Change2 years ago嘉宾来稿:满足中国增长的用电需求 光伏加储能“比新建煤电更实惠”
-
Climate Change2 years ago
Bill Discounting Climate Change in Florida’s Energy Policy Awaits DeSantis’ Approval
-
Renewable Energy9 months agoSending Progressive Philanthropist George Soros to Prison?
-
Carbon Footprint2 years agoUS SEC’s Climate Disclosure Rules Spur Renewed Interest in Carbon Credits
-
Greenhouse Gases1 year ago
嘉宾来稿:探究火山喷发如何影响气候预测











