How AI is transforming the PV industry? Introducing Huawei latest FusionSolar 6.0+ solution
Mayank Mishra, India’s Regional Sales Director, Smart PV, Huawei
With the rapid development ICT technologies such as AI, cloud, big data, and 5G, Huawei has engaged with experts in the field and has released 10 emerging technical trends for smart PV for 2025. There are 3 key directions that are worth paying attention to.
- Full intelligence: in the future, data will become the most fundamental resource for the digital world and all the data in the PV plant from modules, trackers, inverters and etc. shall be detected. We expect 90% of the PV plants will be fully digitalized by 2025.
And the mass data will be processed with AI to help improve the system operation efficiency which is previously performed with only human experience. We expect AI to be deployed in system design optimization for detection, solar storage collaboration, and ultimately making the whole system smarter. By 2025, 70% of PV plants will adopt AI.
We believe unmanned operation and maintenance will be the mainstream. Drones will be deployed for the inspection and assistant robots can help repair and replace devices with remote control. We expect by 2025, 80% of the PV plants will be unmanned at least with light manual operation.
- Grid supporting: currently the installation capacity of renewable energy accounts for 10% of the total power plants where energy generation accounts for only 2% at this moment. In the future, renewable energy become the main energy source as significantly more PV plants will be connected to the power grid. Therefore, we expect PV plants shall transform from following the grid to actively supporting the grid. When the grid integration failure occurs, the PV plant should not be tripped, and the output power should be stable with low harmonics in order to maintain the voltage and the frequency of the grid.
The PV plant storage is expected to be a new trend for grid supporting. In 2025, we expect 30% of the new PV plant will have energy storage. However, the smart collaboration between the solar and the storage to improve the system efficiency will be another new project for the PV solution suppliers.
Also, 80% of residential systems will connect to virtual power plants (VPP) networks by 2025.
- System and the components upgrade: with the development of the material and heat dissipation and algorithm, we expect components in PV solution will be upgraded in 3 ways. Modular design of components for flexible deployment. Second is power density, in the future, inverter power density will increase by 50%. Lastly, the reliability, availability, security, safety, resilience & privacy will become essential requirements.
Based on these understandings, Huawei created FusionSolar 6.0+ Utility Smart PV solution consists of smart inverters, controller for the data transmission and control, the smart transformer station for overseas IPS solution, and Smart PV management system for real-time monitoring and management. Our latest improvements on the solution are testaments of how we deploy AI technology in FusionSolar: 1) Smart Tracker Control Algorithm, 2) Smart I-V Curve Diagnosis 4.0 and 3) Smart Grid-connection Algorithm. At Huawei, we use AI-Boost FusionSolar Smart PV solution aiming to reduce LCOE and supports grid with the following 3 key features.
- Smart Tracker Control Algorithm:
The tracker can significantly increase the energy generation. More and more PV plants deploy trackers to reduce LCOE. Currently, the trackers rely on astronomic algorithm to ensure the modules to be perpendicular to the direct sunlight. However, the uneven landscape in the PV plant is not being considered. The shading between different rows of PV panels could reduce the yields. In cloudy days, the diffused light accounts for higher ratio than that in normal days. And the tracking angle without optimization can’t harvest as much as power when the light is scattered.
Huawei integrates AI and the collaborative energy generation with tracking algorithm to optimize the tracking angle through the recurrent neural network based on the mass data from the plant performance. And during the backtracking, Huawei adopts each row to avoid inter-row shading and capture more solar energy.
What we introduced here is only achieved by the algorithm upgrade without any cost of additional devices. We have tested this technology in PV plants in different latitudes. And the results show, compared with standard mode, the smart control algorithm can improve the energy generation by at least 0.5%. This technology utilised standard interface and communication protocol and it’s quite easy to collaborate with different vendors of trackers.
- Smart I-V Curve Diagnosis
We also use AI to reduce OPEX as we improve I-V curve diagnosis. This idea came about between us and our strategic customer. We developed an online inspection and diagnosis of all PV strings. Compared with conventional solution that uses on-site sampling test and time-consuming manual analysis. The smart I-V curve diagnosis can largely shorten the site inspection time from weeks to half a day. AI learns based on experts’ knowledge and can accurately identify 14 types of string failure which is certified by TÜV. These diagnosis can further provide trouble shooting solution to close loop the O&M process. This function has already been deployed in 5GW PV plant worldwide at the moment and the number is still growing.
Huawei continues to upgrade the algorithm to suit more and more scenario. This year, the 4th generation can support diagnosis of the mono, bifacial and shingle PV modules. It is also compatible with the mix installation of different PV modules in 1 string. The function is provided through licensing and any of our current customers with Huawei Smart PV solution, either for their existing or the new PV plans, are welcome to test it out.
Malaysian 49MW Solar Power Plant is utilising Huawei Smart I-V Curve Diagnosis to reduce OPEX.
- Smart Grid-connection Algorithm
As the LCOE of PV continuously decreases, solar is expected to become the major energy source according to the International Energy Agency’s outlook. In the near future, over 30% of the annual newly installed power plant will be the PV plant. However, with the increase in renewable energy construction, the short circuit ratio (SCR) of the power grid is gradually reduced. We could see the SCR in some countries will be going down quite significantly. Or in another way, more and more regions will face a new challenge of weak grid. It is expected that between 2025 and 2030, the SCR will be lower than 2 in a lot of the countries.
Huawei has been researching on the grid connection technologies for many years and the improved its control accuracy. Currently, Huawei has achieved stable grid-connection in all PV scenarios. We are confident that our algorithm could help our clients to build a robust system with more resilience. The smart inverters could support SCR as low as 1.5 to avoid frequent synchronization failure incidents. Moreover, Huawei solution is capable to provide better POD (power oscillation damping) control, which has already been considered and partially adopted with some of our clients in Latin America.
In summary, as digitalization transforms key industries, we could see the trend of All Things Sensing, All Things Connected and All Things Intelligent happening in the PV industry too. To achieve this vision, smart inverters are key components to reshape the industry. Inverters are no longer just converting DC (direct current) electricity to AC (alternating current), but to control and diagnose the health of the PV systems, ultimately making the system smart and intelligent. At Huawei, our vision is to bring digital to every person, home and organization for a fully connected, intelligent world. This would apply to the solar energy industry. We advocate building an AI-enabled smart PV ecosystem by working with clients, policymakers, upstream and downstream partners to create a fully connected and digital PV system.