Times are changing, and traditional industries must also change.

This is an old saying, and more and more companies are realizing this.

However, what is the right way? What is the right way to do it? How much value can the transformation bring to the enterprise? We still don’t know, and we are also hesitant in the process of exploring new technologies such as AI.

Affected by the epidemic, the need to resume work and production has once again injected momentum into the transformation. Enterprises are deeply aware that only by integrating technologies such as AI and cloud into production, operations and management, will cross-time zone collaborative production and online office be possible in the future, and the company's revenue will be more stable. Therefore, the digital upgrade of various industries around the world, that is, the transition from offline to online, which began in 2019, has been further accelerated.

Against this background, leading domestic cloud service providers and AI unicorns are actively participating in it. Technologies such as cloud computing, 5G, AI, and IoT are gradually being implemented in the industry to help companies transform.

Taking Huawei Cloud EI Enterprise Intelligence as an example, it has already implemented more than 600 projects in more than 10 industries. Recently, based on its in-depth understanding of the industrial field, Huawei Cloud launched Huawei Cloud EI Industrial Intelligence 2.0, with the aim of providing the industrial field with more powerful AI computing power and a user-friendly AI platform. This is rare and of great significance to the industry.

New ideas for industrial AI implementation

In fact, there are many manufacturers both at home and abroad that provide solutions for industrial intelligent transformation, including traditional manufacturing giants and established industrial solution service providers, as well as cloud service providers and new software and hardware suppliers in various niche areas. Relying on their own technical architecture and ecology, the products and technical solutions of these manufacturers have also formed their own characteristics.

Among these many solutions, what is the difference between Huawei Cloud EI Industrial Intelligence?

The vast amount of industry knowledge accumulated over the years in the industrial field can solve qualitative problems very well, but in many scenarios, these mechanism models cannot accurately match the fluctuations of working conditions, and the industrial process is still a "black box". In addition, a large amount of implicit knowledge possessed by factory masters and craftsmen needs to be passed on and replicated.

Jia Yongli, President of Huawei Industrial Internet Solutions and President of Huawei Cloud Artificial Intelligence, said, "For AI to enter the core production system of enterprises, it needs to be deeply integrated with the existing mechanism models and invisible industry knowledge in the industry to release the potential of production factors and production resources. This is the essence and key of the Industrial Internet. As AI technology enters the production system of enterprises, AI will bring significant quality improvements and cost benefits in key production links. I believe that in the next 5 to 10 years, the Industrial Internet will reshape the production model of enterprises."

Figure | Jia Yongli, President of Huawei Industrial Internet Solutions and President of Huawei Cloud Artificial Intelligence

The role and prospects of AI in industry are beyond doubt. However, through continuous in-depth exploration, Huawei Cloud has found that there are some common problems in existing "Industry + AI" attempts, such as the lack of an AI development and operation management platform suitable for industry, the difficulty in developing and sharing industrial mechanism models, and certain instability in production systems that integrate AI.

Industry knowledge is the key to improving quality, reducing costs, and increasing efficiency for enterprises. Huawei Cloud envisions that building new technology architectures around the industry can truly promote industrial upgrading, and this is also the starting point for its service to industrial manufacturers.

Based on existing capabilities, Huawei Cloud EI Industrial Intelligence 2.0 will be born to solve this series of problems.

Simply put, Huawei Cloud EI Industrial Intelligence 2.0 not only deeply integrates industrial mechanisms, it will also have an industrial AI development, operation and deployment platform to help the industrial field quickly implement AI and achieve intelligent upgrades.

There is a difference between deep integration and applied AI

Huawei Cloud EI Industrial Intelligence 2.0 is not just an addition of functions and a simple upgrade, but a change in positioning and thinking. Huawei Cloud believes that for enterprises in the industrial field, Huawei Cloud EI Industrial Intelligence brings not only simple application of AI, but also the deep integration of "AI + industrial knowledge", which is the most critical point.

What difference can "deep integration" bring? Measuring by project results has the greatest credibility.

For Shiheng Special Steel, a large steel conglomerate, the product of the deep integration of "AI + industry knowledge" is the AI ​​intelligent coal blending solution. Huawei Cloud engineers have visited Shiheng Special Steel several times, and have worked with coal blending experts on the coking production line to sort out the details of the coal blending process, and discuss the optional directions, technical solutions, and optimization space in the optimization decision. After repeated discussions, the two sides reached three key consensuses on the coal blending optimization scenario: accurate prediction, collaborative optimization, and continuous iteration. At present, Huawei Cloud's coal blending optimization AI model has supported stable commercial use after several rounds of iterations, with an accuracy rate of more than 97% in coke quality prediction, helping customers reduce the average coal cost per ton of coke production by about 15 yuan. Based on an annual output of 750,000 tons of coke, the raw coal consumption is about 1 million tons, and the company can save about 15 million yuan each year.

For PetroChina, the product of the deep integration of "AI + industry knowledge" is the exploration and development cognitive computing platform. PetroChina introduced Huawei Cloud EI Industrial Intelligence into production practice, using artificial intelligence technologies such as natural language, knowledge graph processing and machine learning to build, calculate and apply knowledge systems. This cooperation helped PetroChina achieve the goal of reducing costs, increasing efficiency, and increasing reserves and production.

Moreover, Dagang Oilfield has realized intelligent identification of oil, gas and water layers with the help of the exploration and development cognitive computing platform, shortening the evaluation time by 70% and significantly improving work efficiency. In the field of oil and gas production, the Internet of Things technology and machine learning methods have been used to realize quantitative diagnosis and remote real-time online management of oil well conditions. Based on this, the accuracy rate of abnormal condition diagnosis has reached more than 90%, reducing operation and maintenance costs by 20%.

Five capabilities support AI implementation

Today, based on Huawei Cloud EI Industrial Intelligence's exploration and practice in the industry, Huawei has condensed project experience into the five capabilities of EI Industrial Intelligence 2.0, thereby promoting AI capabilities to more companies.

In order to enable enterprises to develop industrial mechanism models more quickly, based on powerful knowledge graph capabilities, Huawei Cloud has added a multi-language mechanism model development platform to EI Industrial Intelligence 2.0. Simply put, by turning the mechanism models of various industries into operators that can be recognized by industrial intelligence, Huawei Cloud has built a powerful standard library that allows users to query, call, and share complex industrial knowledge on the platform.

But being able to query is not enough. Huawei Cloud has also built a low-code development platform that integrates industrial mechanisms and AI, and incorporated its own ideas. First, AI models are built based on scenarios and industrial mechanism characteristics, making AI models explainable and more reliable. Second, in order to enable industrial customers to develop models more efficiently, Huawei Cloud has made a "low-code" design, allowing users to complete the integration modeling of industrial mechanisms and AI by dragging and dropping, greatly improving the efficiency of model building.

In addition, Huawei Cloud EI Industrial Intelligence 2.0 has built an industrial-grade AI deployment and operation management platform by matching the organizational structure - main factory - factory area - workshop - production line, but the platform is not only about realizing industrial operation and management functions. Intelligent experiments in the chemical fiber industry have proved that it can be adapted to existing equipment and control systems in a "plug and play" manner without stopping production. Low latency, minimal consumption of storage and computing space, ensuring data privacy through federated learning, and ensuring model security using trusted computing technology are all capabilities of the platform.

A powerful calling library, a convenient and reliable development platform, and a deployment and operation management platform are how Huawei Cloud puts its philosophy into practice.

at last

In fact, building an ecosystem and underlying framework requires taking into account many links and details. For giant companies, even with a strong industrial chain and the richest resources, it is a solid challenge and the road ahead will inevitably be long and difficult.