SmartHelio CEO and Founder shares how AI and automation will be a game changer for the solar industry in future.
Q1 First of all, we’d love to know more about your Background, academics and work experience. How did you start this company?
My name is Govinda Upadhyay. I am the CEO and founder of SmartHelio based out of Switzerland and an office in New Delhi. I grew up in a middle class family in India and did my schooling in Kendriya Vidyalaya.
Then, I graduated from BITS Pilani, India in 2010. I came to Europe for my higher studies in Sustainable Energy Technology from KTH, Stockholm and TU/e Eindhoven in the Netherlands. I am very passionate about Renewable energy and cleantech. I have been listed in Forbes 30 under 30 for my work in sustainable energy and I was also awarded with the prestigious ‘Swiss Ambassador Award’ and European Innovation and Technology (EIT) award.
I did my research from EPFL, Switzerland on sustainable urban planning where we worked on how to meet urban energy demand through renewable sources with respect to increase in ambient temperature due to climate change.
My specific interest was on microclimate analysis and sustainable urban planning. During my research, I realised that we already have the right tools and technology but still clean energy penetration in our society is slow. It was in 2016, that I started my first start-up ‘LEDsafari’ on educating children about sustainability. Our team traveled across countries like India, Latin America, Africa and Europe, teaching school kids about batteries and LED powered by solar, through Lego-style solar kits.
In 2018, while travelling across the world, I realized that solar plants were getting sick and underperforming. Although solar panels are designed to perform for 25-30 years, as solar plants get older their performance declines and they tend to live for only 10-15 years, which is far less than their expected life. Thus most of the solar panels end up as e-waste and also make it less profitable.
Keeping this serious problem in mind, I started ‘SmartHelio (Helios, (Greek: “Sun”) in Greek religion, the sun god, sometimes called a Titan)’ in November 2019, Switzerland to become the doctors of solar plants. We provide an affordable advanced technology to not only help solar asset managers automatically diagnose sick solar plants, but also help them maintain the health of their solar assets so that they can live long and perform well.
The vision is to accelerate adoption of clean energy in our society by using affordable deep tech. I started the company with this vision and today we can proudly say that we are a world’s leading solar diagnostic startup. SmartHelio has revolutionized the approach to solar system diagnostics by using patent-pending proprietary Edge Computing technology running on physics-based algorithms.
The analysis uses real-time data from multiple data platforms irrespective of their database schema, nature and environment. The technology not only detects the problem sooner, but also predicts future problems in the system. Furthermore, necessary corrective actions are automatically prompted.
Our Edge computing based IoT sensors collect data from the modules and analyze it in our cloud to provide real-time solutions and actionable steps for the maintenance staff. We not only localize the fault such as connector failure, wire rusting, panel degradation, shading, weed growth, soiling (40+ faults) etc . but also prescribe corrective measures to be taken to fix the fault.
We also predict the faults so that the solar asset managers are able to decrease the downtime of a solar plant and also maintain them well. Our mission is to become the ‘brain of clean energy’. We have started with solar PV, but we have built our technology with a modular approach so that it can be scaled to batteries, wind, hydrogen etc.
Q2 Please tell us in detail about your company and the kind of products and technology it offers. Also, tell us about your team and your mission.
SmartHelio is headquartered in Switzerland with an additional office in New Delhi, India. The team consists of PhDs and industry experts with 100+ years of experience in the clean energy domain. We have our team coming from EPFL Switzerland, Fraunhofer (Germany), Michigan University, IIT Bombay , IIT Kanpur and BITS Pilani.
We are highly skilled engineers and scientists with deep knowledge of solar technology and data science. The company is deeply rooted into research and development to leverage frontier technologies like AI, ML, edge computing and IoT to tackle the problem of climate change by improving the performance and sustainability of solar PV systems.
We have industry experts who have spent a long time in the renewable sector while handling operation and maintenance of solar assets. We are supported by advisors coming from Credit Suisse Energy, responAbility impact fund and Ex CEO of Meyer Burger. Further, we have marquee investors coming from clean energy and impact backgrounds.
Recently, Ex-Global Head of O&M, Tata Solar has joined our team to provide us with more insight into the renewable sector and challenges of the industry which data analytics can help mitigate. Our team is well balanced.
We are a team of 18 out of which eight people are based in India. The rest of our team members are based in Europe and mostly focused on R&D.
We provide our clients with two kinds of services. First is our cloud platform where we collect data from the monitoring portal or SCADA or inverters and start providing real-time diagnostics with root cause analysis of under-performance.
SmartHelio diagnostic platform
Second is our ‘HelioHealth Sensor’ which is like a “Fitbit” for solar panels. These panel-level monitoring sensors are smart sensors. They are not data loggers. They basically measure the current, voltage and temperature at a high frequency of the panel.
Then, they find anomalies in the panel’s behavior at the sensor level itself using edge computing. Next, it sends the relevant/anomaly data to the cloud because we don’t want to use it as a data logger and start sending a lot of data in the cloud. We minimize data transfer to the cloud by 80% using the edge computing capability of our sensors and speed up our fault detection/prediction at panel level.
One interesting thing about these sensors is that you don’t have to put these sensors in every panel. You put in one of the panels in the string and then you work on the reference panel. For example, on the MW scale, you just need 5-6 sensors per MW. Then we use the data from the sensor in combination with the data coming from the inverters (SCADA) to really diagnose better and understand the impact of a panel on the entire string or array.
The sensor usage is multifold and even can replace certain devices in a solar plant. For example, solar irradiation can be calculated using sensors indirectly using advance algorithms, hence the number of pyranometers needed in a solar plant can be reduced. Moreover, You don’t need dust sensors, temperature sensors and even string monitoring at times. We have actually seen our clients use this sensor to reduce their CapEX cost up to 30%, which is quite interesting.
To summarise, we have two offerings- One is our cloud platform and the other is our cloud plus panel-level HelioHealth sensor. Everything goes into the service model. So, you pay per MW per year and if the sensor goes bad, we replace it. Also, for a large portfolio we also have a licensing model.
We are catering to some of the most renowned companies of the world including Tata, BKW to name a few. There has been a constant increase in demand for our smart sensors especially in cities like Geneva where data acquisition and data quality has been a big issue for the solar industry.
Q3 Can you tell us about your financial partners and their role in your company?
We are a fairly young company (1.5 years old) but we already have strong financial partners like Alternative Bank Switzerland, which usually invests in green projects, as one of our lead investors. We are also supported by Venture Kick fund which is basically a Swiss-based Venture fund for deep tech . We have received help from the state of Switzerland in terms of providing us exposure to the energy companies’ network and helping us with operational finances, if needed.
We have received R&D funds from Europe and the Swiss Government to continuously innovate and advance our technologies. Our financial partners are our major support. It is their belief in our mission as a company that motivates us to do better. We are also looking for more investors who believe in our vision to make clean energy more sustainable using affordable deep tech.
Q4 What role will new technologies like artificial intelligence and Blockchain, play in the renewable energy sector?
Renewable sector needs transformation. Today, most of the countries are dependent on fossil fuels for their energy needs, however, things are not going to be the same in future. The energy demand will see a major shift to renewable sources by 2050, with only 20-30 per cent coming from fossil fuels.
The biggest challenge for the world would be to provide a stable source of energy through renewable sources to meet our demand. The biggest bottom line for renewables is uncertainty. We don’t know how much it will produce, given the contingencies in predicting the weather.
We are also not very well equipped to control the generation from renewables. So, renewables cannot provide a predictable baseload to the grid and that’s why we have seen the news in Germany and in the UK, when we have a lot of solar and wind and the grid gets unstable. If we cannot control them, and if we don’t decrease the uncertainty in renewables it will be very difficult to push a lot of renewables into the grid.
That’s where technology like edge computing, automated controls and artificial intelligence will play a vital role so that renewable energy is available on a large scale to meet the ever increasing energy demands of the world.
I believe Blockchain will help individual households to become ‘prosumers’. The reason why I am using this word is because right now, energy systems are located outside urban areas. For example, we have dedicated fields for solar plants, wind farms and the same will be for hydrogen farms also. But, in the long run, we will see a lot of households producing their own energy. Thus, they become both producers and consumers of energy. Thus they would be ‘prosumers’.
This trend is already visible in western countries while in India it is gaining acceptance. That’s where blockchain would come handy because it can help us to enable peer-to-peer energy trading, making a pure prosumer of these clients. They can not only consume power but also produce it in a very confidential and transparent way through the grid. And Blockchain would not only help in making this possible but will also help us to determine the quality of the system.
We have already seen how Blockchain is helping the pharmaceutical industry to identify fake medicines. Blockchain technology is even helping luxury brands like swatch, Rolex etc. to ensure that consumers get original products.
I believe Blockchain would do wonders in ensuring distribution of quality goods like solar panels and other solar equipment. These are just some of the many aspects where Blockchain can play a crucial role.
Q5 And what about cloud technologies, how are they going to affect the renewable sector?
Already the cloud technology is helping us in forecasting and thus decreasing the uncertainty associated with renewable energy. This helps in optimum utilization of clean energy. With 5G technology, data collection will be easier for us. It will thus, help us to monitor performance of assets in an efficient way. A lot of investors will feel more confident in investing in such technologies. Many asset owners are already pushing their data on the cloud.
Q6 What solutions and offerings do you have for different market segments such as utility-scale solar, Rooftop Solar, Energy storage & Electric vehicle, Off-grid solar etc?
Currently we have two offerings. One is the cloud technology where you give us your solar data and we can diagnose it. The other one is where we provide smart sensors that can be implemented on utility-scale or rooftops and in off-grid areas. Recently, we also did battery analytics using our sensors.
Right now, we are working on how to make our technology compatible for Electric Vehicles. We have a R & D partnership with research institutes in Europe and are currently focusing on strengthening our analytics on different clean energy sources. I can’t talk more in detail about our R&D but I can assure you that something disruptive is coming soon.
Q7 Please tell us about the role of data and analytics, and the tools and their role in asset management operations and maintenance.
I think there are three parts to it. First requirement is that we should be able to collect data as precisely as possible. The second is that we should be able to know whether this data makes sense or not. For this, there has to be an intelligent algorithm to ensure high quality of the data.
Also, at times, we miss data points, so our artificial intelligence predicts those missing data points. And the third one is, of course, analytics which is not just about comparing two strings (whether a string is working or not), rather it should tell us the root cause of a problem and how to fix it. It is about diagnosis, prediction and prescription.
Presently, in the solar industry, all the monitoring companies are providing analytics. This analytics, however, is limited to comparing one solar inverter with another, and if some strings/arrays are not performing, then they send a person to localize the faults and fix those strings/arrays.
Now picture this! What if a company is able to know exactly why a fault has occurred. What if a company is able to know whether the fault is due to soiling, connector failure, wire rusting or ground fault? This could not only minimize the maintenance effort, but also save huge costs incurred on operation and maintenance of solar assets. Also, it will bring transparency and reduce human error. Also, imagine we are even able to predict these faults much before they occur.
For instance, if solar companies are able to know why a fault has occurred and what would be the effect if we do not fix it immediately, they can then utilize and prioritize their resources accordingly.
After talking to a lot of operation & maintenance colleagues from different companies, I came to know that still a lot of plant’s data was maintained on excel sheets. That’s insane and I was surprised by that. All the companies are moving towards monitoring portals, but I think the future will be more diagnostic and prescriptive. Because just monitoring does not help in maintenance of the health of solar assets.
Q8 Tell us about the data logger, its features, advantages and the value it offers. Please let us know about the sensor, its features, advantages and the value it offers.
First of all, it is not a data logger, it is a smart IoT sensor that does edge computing, which is possible due to the intelligent processor that our sensors have. These sensors already have algorithms running and they can tell if a panel is experiencing shadow, PID, connection failure or bypass failure etc.
Our sensors are smart because they understand the health characteristics of solar panels very well. They communicate with our cloud directly via Wi-Fi, 3G or 4G network. They have bi-directional communication so the cloud always ensures that the sensors learn better about the systems over the time. This was made with the purpose to ensure quick on-ground decision making possible for the monitoring staff of the company.
In the end, our aim is to control the systems in such a way that decision making is easier and faster for the solar companies. Most of the data loggers just measure and send the data to the companies. At times, they do not even measure the data, they just read the data from the inverters and send it to companies.
What we wanted was to fundamentally know the actual values of current, voltage and temperature of the panel and what factors are affecting the overall power from a panel. That is the aim of this smart sensor. It is called HelioHealth, “fitbit of the solar panel”.
In 2022 we would be releasing the next version of our smart sensor, which would be called ‘HelioEdge’. It will do mesh computing. To put it in simple words, these sensors would be able to communicate with each other so as to make the solar asset maintenance more efficient and easy.
The third version of the sensor would be called ‘HelioControl’ and will be launched in four years’ time. HelioControl will be more advanced than its predecessors and will be fully automated so as to know when to switch off and switch on the renewable source. We believe that HelioControl, once launched, will enable solar plants to achieve a virtual power plant state because we would have much better control of our systems.
Q9 Kindly let us know about your clients and some case studies and key learnings.
Our journey has been very interesting so far. We started with off-grid clients when we started with Mera Gao power. Then, we worked with companies like Cleanmax, Amplus as our clients. Slowly we moved on to cater to bigger clients like Hinduja, Tata, Adani.
In Europe, we have clients like BKW- one the largest energy companies in Europe. The City of Geneva uses our technology to manage their solar assets. There are multiple things that we learned in this journey so far.
We realised that we need to make solar data analytics much more convenient for the companies in a sense that all the analysis is easily understood by asset managers.
In the process, we also have to ensure that the information is translated in a very simple language to the person on the ground. We are working on making sure that the message from the analytics comes in a way that people can understand and take quick actions.
What we learned was that clients trusted us and they started giving us a lot of use cases. We have clients saying, “I have a Solar plant system which has combiner boxes but they have 20 strings together, we do not know what’s happening at a string level.
And, there are inter row shadow, shadow from growing vegetation or they have a connector problem and all. Can you help me to really know which string has which kind of problems?” This is very difficult because you are measuring data which combines 20 strings at the combiner box level. So, we are constantly working with them to determine how with one single point (combiner box level data) we can determine something out of it.
It was an extremely difficult problem but the client trusted us and now we are working together. They share a lot of knowledge with us. Also, it was our clients that gave us a new business model. They said, “If you combine your forces together with some monitoring companies it will be a boon for all.”
However, this is very difficult. When you talk to monitoring companies they have their own offerings but they do not have all the features that we provide. I will be more than happy if we could combine the two offerings. And we are in fact working on it.
We are now in collaboration with multiple monitoring companies where we just plug our APIs with monitoring platforms and give it to our customers. So far the feedback from our customers has been great. Customers are very happy.
Infact, one of the customers said, “Initially I didn’t even know that this kind of analysis exists. I used to send a maintenance guy to do the ground check of solar plants. But now, I am able to see the issues in real-time and also have an insight on the reason for the occurrence of the fault.”
They learn also. They like that we are able to teach them when any fault happens, because prior to our analytics, they did not know if it is shadow, connector failure, bypass failure or soiling.
One client told us that some of the faults took very long to be identified because it was random. It was only after the analysis, they were able to clearly identify the kind of faults that were happening. To them it seemed like magic, but for us it is pure science.
We strive to make asset management much more convenient and cost effective for all the leading companies that we are working with. We help asset management and O&M staff to diagnose the fault, so that the performance of the plant is highly profitable for the company. This way, we can push renewable and solar much faster into our society.
Q10 Please tell us about some of the awards that you have won.
We have won around 20+ global awards and recognition since our inception in late 2019. Recently, we were selected for the Amazon AWS Energy Accelerator. We were one of the top 10 companies selected for this program. During the program we would learn from AWS how to create cloud architectures that could take all the data and process them. Also, we would also get access to their clients.
A lot of companies are shifting to renewable energy such as Shell, EnI, Galp etc. With big giants like Reliance and Adani Group shifting to renewables, it will be a game changer for India. We are already working with big energy companies who are undergoing energy transition and they need new approaches to renewable energy. We are happy to be able to help them do so. We are also labelled by the Solar Impulse Foundation which have added to our credibility.
Recently, we were awarded Start-up of the year by Solar Power Europe. We also got a one million grant to further work on our Edge computing sensors. SmartHelio has won multiple awards in just two years and has been selected in various incubators globally.
To name a few: We were awarded with Solar Impulse Label, Startup Of The Year By Solar Power Europe, Venture Kic, Company Of The Year (Data Analytics) By Solar Quarter, Get-in The Ring Impact Swiss Champion, Energy Tech Challenger Europe AI Category, Energie Venture (Canada), Aws Energy Accelerator (Seattle), Masschallenge Switzerland, Climate KIC (Switzerland), Exel Accelerator (Sanghai) etc.