Kazuhito Yokoi: At Hitachi, we’ve gone deep into some AI projects where, using generative AI, our factory engineers can develop and generate custom code to get final calculations from massive amounts of data. Before the introduction of generative AI, plant engineers needed to ask IT engineers to direct the development of entire systems, but now, plant engineers can create and expand their role in the IT space. Such a shift could shorten the process and time required for innovation research and development.

Bob Gill: The UCS released this time is a new “software-defined” control system that subverts the physical form of traditional DCS. What new value creation do you think such a completely innovative product will bring to the enterprise?

1797-PS2N2 Christian H. Sboro: UCS certainly solves a lot of enterprise run-layer problems. For example, PGN has a lot of storage and monitoring equipment related to natural gas pipelines. If abnormal conditions such as leakage or corrosion occur, personnel are usually required to go to the site for on-site inspection, which actually has a certain delay. The emergence of UCS can monitor the factory equipment in real time and preventatively, so as to avoid the occurrence of safety accidents and parking accidents. Most importantly, UCS enables predictive modeling and action based on future control parameters.

Kazuhito Yokoi: Hitachi had a lot of edge devices, and it was very difficult to manage them centrally. UCS’s ability to capture and analyze data from a single cabinet is impressive. I think UCS is a very promising solution for the vast majority of people to enhance the overall performance of the business.

Cui Shan: It is very clear that the emergence of UCS can bring significant cost reduction value to the enterprise, about 80% of the cable cost will disappear. Secondly, UCS breaks the barriers between traditional OT, IT and AT, and all applications can operate optimally independently with just one click, which is not only the optimization of labor costs, but also a breakthrough innovation for enterprises. At present, UCS has achieved commercial application, and the central control technology is expected to support more enterprises to achieve digital intelligence value breakthroughs.

1797-PS2N2

Bob Gill: Taking the process industry’s first time-series large-scale model TPT, which was released today, as an example, combined with the actual needs and pain points of their respective enterprises, please talk about the typical value and application prospects that generative AI technology may bring to their respective enterprises and even the industry.

Brad Lee: EMQ has been thinking about how to link the physical world and artificial intelligence through software. Our cooperation with CTC on UCS is to realize the transfer of log data between machines through our software. I think the prospect of generative AI is very optimistic, but it is still in the stage of product productivity development, and the future research and development of deep-seated AI as a product itself needs to be continuously explored.

1797-PS2N2 Cui Shan: TPT undoubtedly opens up a new way for plant operation in the process industry. First of all, TPT creates a new mode of supporting multiple application scenarios through one software, which can help enterprises save a lot of software costs. Secondly, through the acquisition and training of massive data, TPT has shown amazing adaptability across devices and working conditions, providing enterprises with more inclusive and more reliable production optimization. With TPT, we believe we can make more of the traditional industrial impossible possible.

Christian H. Sboro: As a petrochemical company, PGN really needs predictive maintenance to ensure the stable and safe operation of its equipment. PGN currently generates time series data, but these data are only recorded as logs and are not fully used. With the advent of TPT, our time series data can be further analyzed to make more accurate predictions about the device.

Bob Gill: The accelerated implementation of AI technology in the future needs to strengthen ecological construction in terms of industry and technology. How to build AI ecology? How do you practice it in your company or field?

Kazuhito Yokoi: The AI ecosystem has some similarities to the open source ecosystem. In the construction of the open source ecosystem, we define the software architecture, and then complete the function implementation with code. Through daily activity communication, new functions and new knowledge are shared among ecological partners. I believe these lessons can be reused in the AI ecosystem.