Green and Innovative R&D：

Innovative technology development must meet the requirements of environmental protection stipulation, and the R&D are encouraged to strive for the reduction of resource consumption from the environmentally friendly perspective.

• Development and design stage：Remove toxic substances from the environment and avoid residual substances contained in products or polluting the environment.
• Production stage：Reduce the loss of energy and resources and the emission of harmful substances.
• Product inspection：For newly developed and produced items, we conduct third-party inspection in accordance with customer’s specifications.

Regulation compliance：

Procured technologies are the technologies developed in conformity with relevant regulatory requirements.

Promote smart logistics and services：

Build a real-time database system (PI), actively incorporate new elements of "Internet of Things" and implement smart logistics and services.

Cultivate AI management talents：

Cultivate a new generation of AI management talents and create new value.

AI for catalyst efficiency optimization – EOG Plant

By way of the further process and deep learning of AI through the experience-based operation data cumulated over the years, the optimized catalyst operation data provided by AI shall benefit future operation with effective succession, promptly and precisely locating the optimized point for catalyst, saving the cost of raw material.

AI for process reaction optimization – Process Development Dept.

The AI established model may help the researchers to quickly converge variables to seek the trends, improving R&D efficiency by lessening the number of experiment groups and saving cost, which differentiate the current formulation ratio adjustment of each batch through multiple experiments and experiences.

AI for new product process optimization – Material Development Dept.

By way of AI deep learning to converge variables and forecast the trends of process parameters. It helps not only shortening new product development timespan, but reducing carbon emission by lowering the production energy consumption.

AI for preliminary maintenance of GAS compressor – Maintenance Dept.

It aims to discover the abnormality rates of equipment by the AI model established through the deep learning from compressor vibration data to predict the key factors that trigger the abnormal vibration and future maintenance timetables, and check for instant maintenance when the abnormal factors appear unstable.

AI for customer’s order forecast – GAS Business Dept.

To optimize production line schedule and number of dispatch vehicle to avoid the likely non-loading dispatch rates by integrating AI deep learning model and customer’s storage tank level gauge data, usage and delivery frequencies to deduct the forecast model of customer demand.

AI for deduction of reactors’ molecular weight parameters – EOD Plant

The purpose is to set the raw material intake amount according to the theoretical amount during the batch production operation, and conduct sampling and testing during the reaction process. If there is a shortage, additional raw materials will be added until the specifications are met before entering into the next stage. Through the establishment of AI deep learning models to predict whether the product molecular weight meets product specifications, it is expected to reduce the number of samplings (or even eliminate the need for sampling), shorten reactor idle time, and improve product quality.

AI for production conditions optimization – EC Plant

The purpose is to establish an AI deep learning model for the production parameters of HPEC & semiconductor CO2 to find out the factors for causing unstable quality. Such applied manufacturing process can stabilize operating conditions and methods, gradually reduce the work duplication and emissions, and stabilize product quality.