Source: Dr. Octopus

At present, intelligent manufacturing has gradually evolved from equipment level intelligence to factory level intelligence, ultimately evolving into platform and data intelligence. The introduction of industrial internet and artificial intelligence (AI) has also become a trend. This is particularly true of the lithium battery industry chain. As a complex system of four major disciplines: mechanical, electrical, thermal, and chemical, as well as a long, volatile, and highly complex industrial chain, it is an excellent battlefield for the industrial internet.

At the 3rd Yanzhi Intelligent Electric Vehicle Annual Conference held recently with the theme of "Cohesion and Innovation for the Future", Zhou Xiaoyi, Deputy General Manager of Dr. Octopus Intelligent Technology (Shanghai) Co., Ltd. (Dr. Octopus), said in an exclusive interview with reporters: "As a battery manufacturer, it is our responsibility and obligation to monitor the battery throughout its life cycle. Dr. Octopus applied AI and Big data technology to the entire life cycle of battery manufacturing and service, bringing innovative solutions to the entire industry.".

Octopus hatched from honeycomb energy

In recent years, the new energy industry has developed rapidly, and lithium batteries, as core storage technologies, have been increasingly applied in fields such as new energy vehicles, energy storage, and data centers. At the same time, high requirements have been put forward for power density, conversion efficiency, reliability, and so on. In this context, it is urgent to use intelligent production methods to improve production efficiency, reduce production costs, and ensure the safety of batteries from manufacturing to use and recycling.

Once upon a time, production lines could only achieve functional availability, supplemented by some simple data collection and monitoring. As a battery manufacturing enterprise with latecomer advantages, Honeycomb Energy Vehicle's standard AI intelligent power battery factory was put into operation in 2019. It also began to explore the use of industrial internet to empower manufacturing, continuously depositing models and algorithms on the platform, and reusing these knowledge and assets into the manufacturing process, improving production efficiency and quality level.

Without the support of AI+industrial internet capabilities, it is difficult or even impossible to accurately analyze the chemical changes and process quality control level of batteries in aspects such as homogenization, coating, electrode production, assembly and welding, and chemical composition and volumetric analysis. For example, in key links such as battery pre charging, K value screening, and capacity prediction, micro trend analysis must be conducted through industrial Big data to achieve real process control and enhanced interception.

Just when AI+industrial Internet technology has become a necessity in the industry, Honeycomb Energy Shanghai R&D Center, which was established in 2021, was renamed Dr. Octopus, aiming to develop AI+Big data edge intelligent systems for energy interconnection and industrial intelligence. Similar to the octopus, which scientists refer to as the most magical species on Earth and an extraterrestrial organism.

The realization of AI+industrial Internet can not be separated from Big data. Zhou Xiaoyi believes that from the perspective of the whole industry, the application of the battery industry is still unsatisfactory, and the polarization is serious. The top enterprises are digging into the application scenarios of AI+industrial Internet, while some waist enterprises still stay in the traditional production line automation.

The practices of leading enterprises are also different. Some enterprises only use AI+industrial Internet to build barriers and Moat. Honeycomb Energy, on the other hand, upholds an innovative and open corporate culture and advocates the establishment of an open alliance from the beginning. Dr. Octopus was established in August 2021 and the AI Bee Energy Alliance was established in December.

The reason why we want to establish an ecological alliance is firstly because the battery industry chain is very long, the process is very complex, and there are many scenarios that our family cannot do. Secondly, our strengths are the determination and mission to have scenarios, cases, data, experts, and long-term optimization. However, we lack platforms, professional algorithms, and cross industry experience, so we must strengthen the alliance

According to Zhou Xiaoyi, the mission of the AI Bee Energy Alliance is to promote innovation and upgrading of the manufacturing model in the new energy industry. Currently, there are over 80 signed partners. The goal of the alliance in 2025 is to achieve a high value-added business model of win-win cooperation, mutual assistance and symbiosis; 3 AI intelligent factories; 10+AI industry application standards; 100+invention patents; 500+AI industrial apps;

Open up double cycle data intelligence

According to Zhou Xiaoyi, Dr. Octopus's business segments include energy interconnection and industrial intelligence, aiming to create a double cycle data intelligence of "health gene" and "manufacturing gene" of power batteries. Energy interconnection mainly focuses on the application of vehicle end and energy storage BMS technology, including control systems related to energy storage stations. The industrial intelligence sector is divided into industrial hardware and industrial software, with hardware focusing on motion control and industrial vision; The software is mainly edge computing, mainly including AIOT platform and AI Big data algorithm, and finally forms an overall solution combining hardware and software.

Zhou Xiaoyi believes that data is crucial for the role of AI+industrial internet, as it can do at least three things: firstly, it can be seen at a micro level and understand the operation of devices through interconnected data; The second is to use microscopic data to discover the root causes and internal laws of equipment problems; The third is to find more accurate solutions and form an enterprise knowledge base based on these reasons.

In addition to the battery manufacturing process, data also plays an important role in the battery usage process. By utilizing the same mechanism, charging and discharging strategies, SOX algorithms, cloud based warning, and other management methods can be implemented for in-service batteries by returning data such as current and voltage. We currently monitor the battery data of over 400000 vehicles, covering services such as safety, performance, early warning, and residual value assessment

He admitted that although the role of data is crucial, connecting the data of manufacturing and service processes can achieve traceability and feedback, identify which links in manufacturing have problems, and help improve product quality, there are also some challenges. Firstly, is there sufficient data and clean enough data for standardized modeling? Only models that are universal enough can be used by everyone; Secondly, in intelligent manufacturing, the current data definition, consistency, and quality are not satisfactory, especially in the standardization of lithium battery processes and products, which still have a lot of room for improvement.

In order to achieve innovation, we have laid out a "hard platform" for industrial internet and some corresponding products. We have implanted control and visual terminals into equipment, BMS terminals into battery packs, and used them as carriers to combine with algorithm software and hardware to form a complete solution He said.

Using AI+industrial Internet and edge computing, Dr. Octopus innovatively connected battery manufacturing data and service data, and realized data intelligence of true two-way complementarity, two-way circulation and two-way evolution.

Building an AI+Industrial Internet Hard Brand

Regarding the development of AI+industrial internet in the future, Zhou Xiaoyi believes that there are three major trends: firstly, informatization and digital interconnection. The industrial internet will move from an intelligent base for enterprise transformation to an integrated application in the industry. It must not be the intelligence of a single system, but the intelligence of the entire industrial internet platform; Secondly, the industrial internet surnamed "Gong" must be combined with industrial scenarios and industrial expertise, as different industries require different models and algorithms to adapt to this industry; The third is to leverage the integration and application of some models and algorithms throughout the entire industry chain, and collaborate upstream and downstream to develop corresponding algorithms on the platform, bridging application scenarios.

In Zhou Xiaoyi's view, the industry does not lack large companies that specialize in AI, but rather companies that deeply integrate AI with the industry. In fact, in the field of industrial internet, internal technology companies or business teams like Dr. Octopus, which rely on a large party or a certain track, not only have some model platforms and cloud resources endowed by Huawei, Baidu, and others, but also can develop corresponding technologies based on the mechanisms of industrial sites. This is crucial for the true implementation of AI+industrial internet.

Now, Dr. Octopus is also collaborating with some leading AI and industrial internet technology companies to supplement his abilities, However, "our advantage is first and foremost that we have a better understanding of industry, as we have been constantly working on industrial sites. We also have a better understanding of Party A and how AI+Industrial Internet can be used and integrated with existing systems and equipment. We also have a better understanding of hardware, equipment, and processes, only in this way can we achieve accuracy standards and response speed standards," he said.