As the solar industry has entered a new age of digital maturity, new challenges are knocking on the door. In a recent webinar hosted by EQ Magazine in association with SenseHawk, various solar industry experts came together under the same roof to share their experience and knowledge as well as look at the possible existing challenges of digital solar asset management. The topic of the webinar was “Digital Tools and Analytics for Efficiency in PV Operations,” and the panel was blessed with maestros like:
- Mr. Rahul Sankhe, President and Co-founder of SenseHawk
- Dr. Negi from MNRE,
- Mr. Prashant Kumar Upadhyay, Manager BS from Solar Energy Corporation of India,
- Mr. C. Chaudhary, COO of Amp Solar,
- Mr. Kapil Kumar, Head O&M division from Azure Power,
- Mr. Dinesh, Head Asset Management from Sprng Energy,
- Mr. SK Madhusudhana, VP and Head O&M, SB Energy,
- Ma’am Shilpi Dangi, Senior Manager from Sembcorp Green Infra Limited,
- Mr. Brajesh Kumar Jha, DGM Distributed Solar Operation and Maintenance from ReNew power,
- Mr. Sudhir Pathak, Head Central Design and Engineering from Hero Future Energies, and
- Mr. Premchand, General Manager at Tata Power Solar Limited.
Watch the excerpt
Mr. Brajesh conveys his take on digital asset management challenges by saying, “While we get lots of data from multiple devices, what matters is whether or not we are trying to utilize it in a proper way. So the key is to do an intelligent digitization. For instance, if there are 100 inverters at site and from each inverter I am getting 10 parameters, it sums up to almost a thousand data for inverters. There should be an automatic system which tells me which 5 out of these 100 inverters are performing low. With this indication, my site team and the technician engineers can go to the plant the very next morning, and hit that particular area. While talking about 100 megawatts, which is almost 400 acres of land, manually monitoring and taking an issue at a particular string level or inverter level is a real challenge. So digitization is necessary to take care of these uncountable missed out losses. On 100 megawatts, if 20 strings are down, it is hardly 120 kilowatts. I’m not going to work about it as looking at the overall perspective and generation, this meager a loss is not going to make much real sense in a day or a year. Hence, automation is a must.”
Being in line with Mr. Brajesh’s views and taking over, Mr. Kapil begins, “One of the biggest existing challenges for the entire industry is assessing the quality of the data. While we have the SCADA system on the ground and the data loggers on the side, are we really getting the right quality of the data for further detailed analysis? Although as an industry, we are already more than a decade old, I don’t see a single company talking about delivering greater value of predictive maintenance using machine learning or artificial intelligence. I have a plenty of data, but are there right companies existing in the market who could analyze it? The lack of sanitization of data, standardization of the equipment, and not much work done on predictive maintenance, are some of the issues we are facing as an IPP and are trying to resolve them at Azure. We have in-house developed a life monitoring system, which can monitor each and every site. All of us still need to look into working on the existing issues.”
Intelligence does not Come Easy
From the perspective of an IPP, Ma’am Shilpi shares what they are internally doing at Sembcorp, “When your assets are growing we realize that it becomes extremely difficult for you to manage all of them. The need of the hour is a centralized monitoring system, with some intelligence. You cannot have all human touch points to give you a reporting, dashboarding, analytics. Once you take in all the data, in the first layer, you utilize it to make it into a very sensible data. Post sensitizing the data, you determine the point(s) for advanced analytics to be performed, as doing it for each and every data you’re collecting and at each and every frequency the data is being collected does not make sense. The level at which you want to collect the data is also very important, alongside choosing the components you want to take, and the components you want to leave out, to ensure you are able to bring out intelligence in the system.”
“Adding intelligence to the system, you will see numerous gigawatts of solar plants already installed up and working, but the biggest mystery of predictive maintenance still prevails. Though we have a list of things to be done on a day-to-day basis, with a lot of data coming in, we are not able to understand how we can use this data to take corrective actions before a plant or a component goes out. Multiple platforms are collecting data from the past 4-5 years at each and every level right from the string to the HT metering panels. However, the output, in the terms of all this collected data, is not on the platform. In essence, to understand whatever data we are taking up with the help of subject matter expertise, the industry definitely needs to move ahead and develop platforms with intelligence incorporated. Furthermore, a huge problem is that you are left on the onus of a couple of forecasting software to CAs, and there is nothing you can do whenever their predictions are going off the track. So, no matter the amount of data being collected across all the phases up till the end point, there is no intelligence at all,” adds Ma’am Shilpi.
The Plan of Action
She continues, “We’ve tactically amalgamated the IoT and the operations department as I strongly believe these two sections should become the core of the system eventually. Once the project is done, the engineering and construction teams are a part of the project lifecycle for a very short tenure. It’s just the asset management team which is hanging on with the project for a really long time. So, whatever needs to be done in order to make the available data intelligent for them, needs to be brought out. We have multiple platforms where we can record, view, monitor, and report in all forms of dashboards, however, intelligence will still take a couple of more years to come into once the quality of data is already sorted. You should not look at asset management as a cost center or an activity which needs to be limited by a budget. This needs to be blended in with the day in day out activities, right from the engineering to conceptualization to commissioning stage, and thereafter. Asset management must be part of the entire lifecycle so that they can add their value and onboard all the documents under the same umbrella.”
“We have a lot of cloud storage, where we are putting in the humongous amount of data, but how effective is the linking of this information? Suppose we have to search for a warranty file for an HT panel today. You may have that data, but unfortunately, who and how it is to be linked, that chain is still missing. With that, I would again like to harp on the point that a lot of data is still there, but the industry needs to mature upon the linking and try bringing out more intelligence out of it ensuring it is actually of use, whether in improvising on the man hour usage, giving us meaningful data-driven decisions or ideology, taking things ahead, and making sure that we are not just achieving what we want, but also have a scope to over achieve what we are now looking at,” concludes Ma’am Shilpi.
SenseHawk has built a cloud-based platform with integrated applications, and a strong focus on data intelligence, quality, and contextual attribution. We support the entire lifecycle, from design and construction to operation and maintenance, of a solar plant. Contact us for a free consultation.