| Summary: | 碩士 === 長榮大學 === 職業安全與衛生學系碩士在職專班 === 107 === Two major petrochemical companies in Taiwan established Big Data Center, Artificial Intelligent and Industrial 4.0 Research Center respectively and emerged into the big tide of AI as the leaders in Taiwan. They are a leader in the field of artificial intelligence and have played a leading role in the petrochemical industry in Taiwan. However, the petrochemical industry hopes to follow the manufacturing industry's success in moving towards Industry 4.0. What is the most important issue?
Some domestic scholars use the Network Physics System (CPS) program to establish key equipment health prediction models for the petrochemical industry. However, after modeling with big data tools, although very accurate predictions were obtained, further research is being sought. When the detailed cause of the fault was found, it was found that it was impossible to continue analyzing the cause of the fault, because the historical data was too restrictive, so the author deeply studied the modeling development of the fault. Research and comparison of recent research on CPS. The literature summarizes the second phase of the study, and at the same time summarizes the relevant content as the conclusion of this report by collating and interviewing several authoritative individuals who have successfully developed Industry 4.0 in the domestic manufacturing industry.
We conclude the following findings from our preliminary study: First, the new manufacturing of new culture: AI is not so much a technological revolution, more like a cultural revolution, AI is more than learning, active learning, active understanding and proactive problem solving, the priority is to establish a new culture and cultivate employees Curiosity and encourage active adventure and learn from tolerance errors. Second, big data analysis: The basis of big data analysis is the domain knowledge (Domain Knowledge) fully understands the field, only to know which data is needed? Third, data collection and data quality: manufacturers and maintenance vendors must be in hardware and software, Also, the collection, pre-processing and preservation of big data, so most industry and manufacturer contracts need to be revised. Data quality also requires complete non-disruptive film, only to analyze the value of use, including the improvement of correct maintenance and process data, otherwise, there is no complete data and resume, the data does not analyze the direction of the target, of course, there will be no results. Fourth, the complete database: the cloud is just a platform, the cloud must have other communities, other petrochemical plants, and self-factory database, the data will be enough, the analysis will be accurate enough, you need to build a complete database of AI, to further It makes sense to analyze the comparison.
|
|---|
