Midea collaborates with Xinyun to cope with the storage pressure of massive AI visual inspection data
With the increasing advancement of intelligent manufacturing, various links in the manufacturing production process, whether it is precision sensors and equipment on the industrial site or various information systems that run through the entire production process, are constantly generating massive amounts of data.
These data are not only essential core resources in the field of intelligent manufacturing, but also a key driving force for continuous innovation in the industry. How to efficiently collect, store, deeply analyze, and scientifically manage this data has gradually evolved into a core strategy for enterprises to seize market opportunities, improve production efficiency, reduce costs, and enhance competitiveness.
Wuhan Midea HVAC Factory is the second factory established by Midea Group with increased investment in Wuhan after Midea Wuhan Refrigeration Equipment Co., Ltd. It was established at the end of 2020 and focuses on the production and sales of Midea household central air conditioning and COLMO household central air conditioning. Currently, its annual output value exceeds 5 billion yuan.
Our self-developed orange robotic arm dances, AGV vehicles shuttle through warehouses, and AI visual inspection system covers all aspects. 5G empowers the intelligent manufacturing transformation of Wuhan Midea HVAC factory, with intelligent equipment, flexible processes, and digital economy applications quickly taking root here.
Xinyun Industrial Intelligent Manufacturing Storage Solution has played a crucial role in the transformation of the home appliance industry in Wuhan Midea HVAC Factory. Its high-performance and low latency characteristics ensure efficient, fast, and secure storage and management of massive data generated during the production process, providing solid and reliable data support for the intelligent upgrade of factories.
High bandwidth and low latency, helping AI visual detection efficiency soar
Wuhan Midea HVAC factory adopts AI visual inspection system to achieve full coverage of key inspection links, ensuring zero quality defects. With the operation of the production line, high-speed cameras on each production line are constantly capturing a large amount of graphic and image data. At the same time, due to the existence of multiple production lines, dozens or even hundreds of data collection terminals will simultaneously collect data, which puts extremely high demands on the read and write capabilities of the storage system.
Midea has deployed the Xinyun Storage SSP00G series storage, which comes standard with 2 10Gb 10G transmission interfaces and can be upgraded up to 10. The 10Gb transmission interface can also be upgraded to 25Gb, 40Gb, and 100Gb.
Multiple industrial control computers within the local area network are connected to a 10G switch through multiple 10G fiber optic load balancing channels and then connected to storage, achieving smooth and high-speed transmission. The maximum transmission bandwidth of a single device is 8GB/s, which can meet the demanding performance requirements of high concurrency, fast access and write, low latency, etc. for massive graphics, images, files and other data.
Global search, quick query
The high-speed cameras on the production line have a high shooting frequency and generate a huge amount of data. It is necessary to be able to quickly retrieve and query this data in order to discover problems in the production process in a timely manner. At the same time, the accuracy of data is also crucial, ensuring the accuracy of retrieval and query results to avoid misjudgment and misoperation.
The Xinyun data management system supports second level global data retrieval and multi-dimensional flexible queries, such as production area, production line, machine number, date, product batch, etc. Unified query reduces retrieval complexity and significantly improves the efficiency of quality inspection data query. Users can easily complete complex query operations without switching between multiple systems or interfaces.
For commonly used or important files, Xinyun supports labeling or asterisks. After attaching the "defective product" label to the defective product images, once production problems occur or specific defective products need to be retrieved, relevant images and videos can be quickly located, greatly reducing manual screening time.
Machine data, scheduled backup
In the traditional production workflow of Midea, data is first stored on the local hard drive of the machine's industrial control computer. When the local hard drive capacity is insufficient, the production line personnel manually copy the data with the earliest generation date to other storage media for storage.
The Xinyun data management system provides scheduled backup function, which allows for the setting of custom backup strategies and the execution of automated backup tasks on a regular basis. This not only saves a lot of manual backup time, but also fully guarantees the integrity and security of production data.
Paired with expansion cabinets to achieve low-cost and efficient expansion
In order to address the challenge of high expansion costs for existing storage devices, Midea has purchased and deployed Xinyun's SSP00G series 48 disk high-performance unified storage, achieving 800TB of usable space storage.
In the future, the storage system will also support easy capacity expansion by adding expansion cabinets. After a single host reaches full capacity, an additional 120 hard drives can be expanded, with a capacity expansion of up to 2640TB. This eliminates the need for tedious setup and complex debugging by professionals, greatly reducing the overall cost of ownership for expansion solutions.
Xinyun Industrial Intelligent Manufacturing Storage Solution is widely used in leading enterprises such as Ideal, BYD, and United Auto due to its high efficiency, stability, and security. In the future, Xinyun will continue to provide efficient and intelligent data storage management and solutions for the industrial intelligent manufacturing field.