Before the emergence of a new generation of information technology, industrial enterprises have been operating normally for hundreds of years. We should clearly realize that the addition of information technology is more like a catalyst.First of all, we need to make clear what kind of business objectives we need to achieve to make the existing production process, industrial products and management methods better.
In fact, the most fundamental goal of big data to support the business transformation of manufacturing industry is to improve quality and efficiency, and realize intelligent manufacturing system on the basis of automation and informatization.On the basis of intelligent manufacturing, the next step is to build a platform, build an industrial ecology, carry out more effective cooperation with the industrial chain, and realize the multiplication development of industrial Internet.
It is also a typical direction for us to achieve the integration of industrial data, industrial equipment and intelligent services.The first level is the equipment level, which is to improve the reliability of a single equipment, identify equipment faults, optimize equipment operation, etc.; the second level is more for production lines, workshops, factories, to improve operational efficiency, including energy consumption optimization, supply chain management, quality management, etc.; the third level is to cross the boundary of the factory industry to achieve industrial interconnection.
Industrial big data doesn’t come out of thin air. Traditional industrial informatization has been going on. We already have a large amount of data from the research and development end, manufacturing process and service link. The process of industrial informatization has been producing a large amount of data. From data to big data, what industry should consider more is the superposition of data and automation domain data, which is the fusion of two kinds of data.In the era of industrial Internet, we need to include more data from upstream, downstream and cross-border industries.
How does industrial big data become the core power of intelligent manufacturing and industrial Internet
What are the characteristics of industrial big data?We summarize the characteristics as “multimodal, high-throughput, strong correlation”.We have summed up about 130 different types of data in the industrial field, with various data modes and complex structural relationships.High throughput refers to the continuous generation of data, high acquisition frequency and high throughput.Strong association means that the data in the industrial scene has a very strong mechanism support, and the data between different disciplines are associated in the mechanism level, not in the data field.
The analysis and application of industrial big data is not the result of deep learning and reinforcement learning.We need to know the mechanism model and quantitative domain knowledge of the research object, which is difficult to advance on the current basis.We hope to find out the statistical relationship between the input and output of the data, and make up for the uncertain and unclear parts of the mechanism and model, which is the basis of the application of industrial big data.
Business leading, data driving industry development
From intelligent manufacturing to industrial Internet platform, the core is to optimize the allocation efficiency of manufacturing resources by using data and models.
Industrial Internet is not the same as intelligent manufacturing. The difference lies in whether there is a breakthrough in the boundary of data and business.At present, too many people pay too much attention to platform capability, while the real industrial Internet is about ecology. Resource optimization is constantly deepening from description and diagnosis to prediction and decision-making, and from stand-alone equipment, production line, industrial chain to industrial ecology.
The key to the success of industrial Internet lies in how to build a business system and how to cross the border.What determines the development direction of industrial Internet must be business driven.From the beginning, we opposed carrying a hammer and looking for nails all over the world. Now many big data and artificial intelligence companies have this problem.
We need to go deep into an industrial field and build a reliable hammer that can just hit the nail with demand. Business driven and problem driven are the essence of industrial development, not technology driven.The key to clear business and data, evaluate data, and truly implement business landing is the collaboration of three elements: people, scene, and algorithm.
The article comes from the Internet