Grid parity’s hackathon


In the energy sector, Huawei has been working with its industry partners to explore digital practices for smart PV since 2014. Collectively, the companies are promoting artificial intelligence and digital technologies such as cloud computing, big data, Internet of things (IoT), and edge computing to work in synergy for further reduction in PV plant costs across the industry.

The PV industry is embracing grid parity globally. Regions particularly rich in solar irradiation resource such as India, Spain, Latin America, and the Middle East, have ushered in the new era. By 2020, most parts of the world will be able to achieve full grid parity. Advancing the global energy transition is crucial to addressing climate change, and bringing LCOE down for grid parity will be the key element and driving force to a clean energy future across all markets. And Huawei is turning to AI to accelerate grid parity.

Taking the lead
In the intelligent era, innovation has become the core competency among leading enterprises. According to IHS Markit, Huawei has been ranked number one in global PV inverter shipments for four consecutive years. The company attributes its successes to customer-oriented innovations that simplify complex problems through fully digital technologies in its electronics, chips, computing, and artificial intelligence.

“Each inverter of Huawei is a smart sensor. The massive amount of data collected by increasing numbers of sensors optimizes more machine learning algorithms and effectively and continuously trains AI models, thereby enhancing the AI capability exponentially,” says Jeff Yan, Senior Product Manager at Huawei. “This helps Huawei optimize and manage power generation and O&M in a refined manner, which radically changes the traditional management mode and brings a revolutionary impact to the new energy industry.”
Huawei’s FusionSolar 1500V Smart PV Solution has five particular characteristics that support the drive toward global grid parity.

Bifacial PV modules+smart trackers+multiple MPPTs
Huawei abandons traditional astronomical algorithms and adopts intelligent trackers and bifacial PV modules with AI algorithms to maximize energy yields by integrating tracker control, power supply, and communication. According to Huawei, a test conducted on a large PV plant in northwestern China demonstrated an energy yield more than 20% higher when using a bifacial modules+smart trackers+multiple MPPT solution, in comparison with monofacial PV modules+fixed trackers+central inverters solution. AI auto-learning enables trackers and bifacial modules to be optimized using the tracking algorithm, which can provide an additional 0.5 to more than 1% in energy gains, according to Huawei.

Informatization of PV devices for building smarter plants
By combining digital components with AI algorithm controls, the solution enables the collection of high-precision digital information to improve energy yields. PV strings can be monitored with the precision of 0.5%. Big data and AI algorithms are oriented to component-level monitoring and refine management to locate old and faulty devices.

AI recognition+Intelligent O&M
Huawei has further upgraded its Smart I-V Curve Diagnosis. The Smart I-V Curve Diagnosis and AI recognition enable the smart diagnosis for multiple components in various scenarios. Currently, the technology has been applied to more than 3 GW of PV plants worldwide. The application of AI enables automated O&M of PV plants. According to Huawei, one-click scanning of its Smart I-V Curve Diagnosis 3.0 ensures the scanning of all strings in a 100 MW PV plant in just 15 minutes.

Another powerful technology used for O&M is its discreteness analysis. In the past, faults were cleared manually, which was time-consuming. Some faults might be ignored completely. “With discreteness analysis, a PV plant in the Shanxi province of China accurately had 283 faults located in half a day. After more than 20 days, the PV plant cleared almost all faults with the performance ratio (PR) improved by 2.52%,” says Yan, “Traditional PV plants, however, take about two months or more to clear faults before feeding power to the mains power grid.”

Grid-tied stability of PV plants based on AI algorithm control
Huawei has established a precise mathematical model for different types of grid-tied scenarios and PV plant designs, and imported related data by testing the power grid waveforms of steel mills and electrified railways into the model. The mass of data is used to train the optimal grid-tied control algorithm. By doing this, Huawei ensures stable grid-tied power generation of inverters even if they have poor power grid waveforms. “The power quality meets or even outperforms standard requirements,” adds Yan.

Health check reports
Huawei offers PV plant health check reports with customized indicators for controlling the operating status of PV plants in a one-click manner. By analyzing the environment, energy yield benchmarking, loss caused by power rationing, and power loss, Huawei uses big data and intelligent analysis methods to perform a one-click health check and generate comprehensive evaluation reports for PV plants with O&M suggestions.

Operations and maintenance managers can quickly find the causes of low energy yields and laggard PV strings, thereby liberating personnel from time-consuming inspection and data analysis, thus reducing labor costs. “The magic of this feature lies in the perfec combination of expertise and AI,” says Yan.

How do these five characteristics benefit customers? “Based on comprehensive analysis and calculation, the AI-aided Huawei 1500V FusionSolar Smart PV Solution can effectively reduce the initial investment by more than $0.05/W and reduce the levelized cost of energy (LCOE) by more than 7% in China,” says Yan. “This is key for speeding up grid parity.” In 2019, Huawei will use the AI-aided 1500V FusionSolar Smart PV Solution to continuously reduce LCOE and accelerate the grid parity process.

This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: