On June 3, 2021, the Shanghai SNEC Exhibition, known as the "world's largest photovoltaic exhibition", was officially opened, a large-scale and influential international photovoltaic event, and Jushi Technology arrived for the second consecutive year. The difference is that this year, Jushi Technology has brought more innovative product solutions and practical rich cases to debut SNEC2021 with a new look.
In this exhibition, Jushi Technology, with the theme of "Photovoltaic AI Unmanned Workshop Expert", took the lead in solving the problem of the current situation of the industry based on strong AI deep learning and machine vision algorithm capabilities. With the photovoltaic industry unmanned workshop solution, AI testing equipment and photovoltaic AI quality management system MatrixSolar, the exhibition hall was unveiled in the N3-270 pavilion.
SNEC2021
The first perspective of Jushi Technology booth
Xiaobian said that on the last day of the exhibition, friends who have not yet come to the meeting should not miss it.
AI empowers photovoltaics and creates unmanned workshops
Jushi Technology AI helps the photovoltaic industry and launches an AI unmanned workshop innovation solution for the photovoltaic new energy industry. The scheme is mainly composed of two parts: the automatic AI detection equipment scheme and the photovoltaic AI quality management system MatrixSolar scheme. Positioned to provide customers with one-stop unmanned workshop solutions and product portfolios, to solve the customer's last mile production pain points.
Based on AI vision detection, the core product modules are richly expanded: deep learning detection system with self-learning ability, photovoltaic machine learning and quality big data analysis, unmanned intelligent EL · VI fully automatic detection, scale distributed deployment, full process quality management, etc. Among them, the equipment and AI software system can be deployed separately and separately, which can create a new AI photovoltaic unmanned production line workshop for customers, or carry out AI unmanned transformation of the customer's old production line, and intelligently upgrade it to an unmanned workshop.
AI empowers photovoltaics, and innovation and breakthroughs continue
MatrixSolar products contain dozens of deep learning algorithm models independently developed by Jushi Technology, including large-scale deep neural network technology. Not only can it realize AI-driven unattended machine vision inspection, but also realize online "self- " learning. In addition to AI visual inspection, in terms of system functions, MatrixSolar is the first in the industry to realize the digital closed-loop system of photovoltaic quality management, and the systematic functions include EL & appearance defect detection of components/battery strings, unmanned management of workshop equipment, yield management and analysis based on machine learning, etc., to achieve digital quality management closed-loop. Reduce the product degradation rate and scrap rate, and help the process to improve from the source to avoid the occurrence of poor batch size.
MatrixSolar AI Integrated Quality Management System
Based on the strong core algorithm capabilities of Jushi Technology, MatrixSolar has strong compatibility and can be compatible with mainstream photovoltaic module inspection such as half-sheet, tile stacking, stacked welding, multi-main grid, single glass, and double glass. Covers more than 20 common EL defects and more than 10 appearance defects.
Super small sample processing capabilities, for low-frequency defects, only a few dozen samples can provide support for the detection of defects.
At the same time, in terms of detection effect, the detection rate of battery string defects is as high as 99.95%, the overkill rate is less than 1%, the detection rate of component defects is as high as 99.9%, and the overkill rate is less than 5%, which far exceeds the peer level.
On the second day of the exhibition, Wang Zhihua, the technical leader of jushi technology photovoltaic industry, was invited to participate in the crystalline silicon photovoltaic module and quality assurance forum, combined with actual cases, introduced the ai to help the photovoltaic industry intelligent upgrade, and create the core product solution of the photovoltaic industry unmanned workshop.
Jushi Technology and Huawei's in-depth cooperation, Jushi Technology's MatrixSolar deep learning system is based on Ascend's powerful AI full-stack basic hardware and software platform, achieving 30% higher computing density and 10% single-shot inference rate than similar products in the industry.
Jushi Technology and Huawei Technology's photovoltaic solution release and case sharing site
AI empowers photovoltaics, and it is time to land
Up to now, MatrixSolar has a number of cases including a domestic manufacturer of monocrystalline silicon tile components and a domestic manufacturer of polycrystalline silicon and monocrystalline silicon half-chip modules. In the latest customer delivery cases that have been accepted, first of all, relying on MatrixSolar's powerful AI capabilities and rich system functions, MatrixSolar has realized multi-node distributed deployment in a single production line, which has helped customers reduce the quality inspection personnel of the production line by 33 people, saved 80% of manpower, and created real value for customers.
In addition, MatrixSolar's innovative features such as maintenance end-of-life management, personnel and equipment management, and yield thermal mapping analysis played an important role in the delivery of "unmanned workshops". More importantly, without the need to "overhaul the existing production equipment", MatrixSolar can innovate "benefit the old", a single set of software systems can drive the transformation of the entire old production line, to meet most of the quality management needs of the production line, to help the industry towards the real "unmanned workshop".