The data possessed by Shanghai Electric is growing day by day, and the scale, type, and complexity of data managed by it are experiencing an explosive growth never seen before. How to extract from massive data information that is valuable for R&D, production, operation, marketing, and management marks a new way of corporate innovation and transformation. The analysis and mining of industrial data is also a way to implement a "heavy technology and light assets" strategy. The hidden value of industrial data has been identified by players from all industries. They demand much on analyzing existing industrial data and hope to optimize their process to make their products and services more competitive on the market by exploring the value of industrial data.

For R&D and application of AI, Shanghai Electric Central Academe has been promoting the application of AI algorithms to the industrial scenarios through exchange between its massive talents and top colleges at home and abroad. For specific problems in each scenario, AI technologies are combined with the industrial Internet, along with expertise, to explore possibilities of the upgrade in terms of product/equipment status monitoring, operational performance management, decision making and analysis in manufacturing, operation and maintenance. These early efforts have delivered positive results.

Equipment operation status monitoring and performance management system: With advanced technologies in sensor, communication, real-time signal processing, and machine learning, data-driven AI analysis is used to enable real-time monitoring of equipment operating status and real-time assessment of key components' health. Any sign of component degradation will be identified in advance, and change in components' operational performance will be predicted in advance by AI algorithms. That allows timely maintenance or intervention to avoid failures.

Smart decision-making system for manufacturing and operation and maintenance: Based on such technologies as feature extraction, machine learning, and optimization control, modules of scheduling and quality control are developed for decision making and analysis of production plans, operation and maintenance plans, and quality management. That allows smart decision making in terms of influencing factors analysis, process control, and tolerance distribution in manufacturing and operation and maintenance. AI modeling can optimize manufacturing and operation and maintenance to improve efficiency and reduce the manual workload.

In the case of remote operation and maintenance of distributed energy, the power generation and load power is predicted by collecting and analyzing operational data. At the same time, user behaviors around the clock are analyzed to offer advice on the power supply. That helps to increase the efficiency in energy production. For wind farms, wind turbines are connected to obtain real-time operation and environmental data. With big data analysis technology, these data can be utilized to enable applications such as turbine health management, turbine vibration signal analysis, wind power prediction, and offshore wind farm maintenance scheduling. For thermal plants, operational data of equipment (steam turbines, boilers, generators) are tracked in a real-time manner throughout their life cycles. Remote diagnosis platforms are built to support real-time analysis of equipment health status and early warning of abnormal status fluctuations. That allows full-life-cycle lean management and operation and maintenance across the plant and reduction in possibilities of power plant failures. The environmental protection group has received an integrated solution of remote intelligent operation and maintenance. Remote video monitoring system and intelligent operation and maintenance information service system are provided for power plants and hydro plants. The systems cover major processes, electromechanical equipment, power and electricity, and other aspects. Modular and component-type management can support life-cycle monitoring of units of all types, and process monitoring and optimizing. It also supports decision making with respect to assessing, monitoring, and predicting statuses such as safety, economy and reliability, and equipment usage and maintenance. AI and data analysis algorithms are utilized for real-time monitoring and analysis of massive hydropower plant operational data, both real-time and historical, to build a smart, integrated solution for hydro/power plants.

While using AI technologies to serve the industrial group, Shanghai Electric Central Academe has been improving AI industrial application solutions based on common demand analysis, technological and experience accumulation, market prospect research, and demonstration engineering applications. With a positioning change from R&D to product promoting, it has established a number of product workshops and a technological translation center to incubate more successful products, and drive productization powered by technological accumulation.

With a strong footing in Shanghai Electric's energy equipment industry, Shanghai Electric Central Academe has developed a series of smart application software and products based on AI algorithm modeling such as industrial data interface and management, industrial data mining and analysis, equipment operation and maintenance monitoring and operation optimization, by connecting smart equipment, control systems, and sensor systems in different fields, and capturing and analyzing massive industrial data generated by machines. In-depth analysis and mining of key business data allow predictive analysis and decision-making. It helps industrial enterprises use a big data mindset to make their products, manufacturing process, and operation smart, find the best path to AI technologies and industrial big data, and offer one-stop service for intelligent manufacturing.

Shanghai Electric Central Academe has updated its positioning and established an AI Research and Application Lab. To explore productization opportunities, it has established product workshops to translate its technological accumulation into real products. With a focus on value and product, it is picking up its paces to form a new R&D structure of sustainable development. By cooperating more with AI institutes both at home and abroad, it has been building its R&D capabilities in AI. Aiming at AI application in the industry, it retains market sensitivity and keeps applying AI technologies to new fields.

The next-generation AI applications will bring great changes to all industries, and they are a must for upgrading to a new industrial system. But to make rational use of the advantages of new technologies and avoid blind innovation, careful attention should be placed in application analysis, technology discussion, and demonstration and guidance. Technical means should be closed combined with business model reforms, and initiative should be taken to explore new models of data-driven development and adopt new technologies to promote enterprise transformation and reform.