The intelligentization wave is sweeping over the world. Data has become an important driver of growth for future manufacturers. However, for most companies, the cloud, AI and big data are all new technologies and concepts, causing frustration in the implementation stage. Google Cloud Platform (GCP) is the world’s leading public cloud brand; it has been cultivated for many years in various professional fields, and it already has corresponding solutions for the manufacturing industry. iKala is a long-term partner of Google Cloud and has a solid foundation for pushing GCP into the applications of different fields.
iKala and Google Cloud work together to promote the implementation and application of the smart manufacturing cloud.
The “Data-Driven Cloud Factory and Smart Manufacturing Cloud AI Seminar” recently held at the Hsinchu Ambassador Hotel gathered experts from both Google Cloud and iKala to give exciting presentations on cloud and AI technologies applied in manufacturing systems. The entire event focused on three major subjects: AI, hybrid cloud architecture, and data analysis.
Alex Lin, executive advisor to directors of iKala, stated in his opening speech that iKala is an important partner of GCP in Taiwan; they are highly specialized in this technology. Enterprises have gradually pushed digital transformation in recent years, and GCP became the main choice for various types of industries to build their future operation architectures. In order to help clients complete smart systems successfully, in addition to being an agent for GCP, iKala also invests in the R&D of AI technology itself, hoping to use the integration of the two major technologies, cloud and AI, to implement their vision of intelligentization.
Steven Shaw, general manager of Google Cloud Taiwan, stated in his speech that since the beginning of the twentieth century, the world has been through several major changes, including world war two, the 911 incident, and the financial crisis of 2007-2008; however, each of these changes also drove the transformation of the manufacturing industry and provided the society with more complete production mechanisms. The COVID-19 pandemic that began in 2020 will also prompt the transformation of the industry. In the future, decentralization, high protection of intellectual property and system heterogeneity, etc., will also become new norms of the post-pandemic era, and cloud platforms will be the best IT choice for enterprises to face these new norms. Google Cloud has also responded to market changes and user needs, continually optimizing the GCP architecture to provide optimized services for various types of enterprises.
For the introduction of smart technology into the manufacturing industry, the event started with the topic “Early deployment of smart manufacturing AI cloud”, carrying out in-depth analysis of the Google AI platform and MLOps and introducing their predicted application cases. The presentation stressed that the traditional practices of repairing and scheduling maintenance after malfunctions have occurred no longer meet the needs of manufacturers nowadays. In the future, AI machine learning algorithms must be used to detect the various operation waveforms of equipment in advance to accurately analyze the failure position and time of equipment in order to prevent unexpected shutdowns of production lines. For example, there were water companies and oil prospectors that used the Google AI platform and MLOps to detect the values of their equipment in order to understand the statuses of the machines and pipelines; this increased the availability of the system significantly and optimized equipment efficiency.
Another highlight of the event was the focus on the hybrid cloud architecture for smart manufacturing, when the new hybrid cloud standards Anthos and Hybrid AI were introduced. Among the hybrid clouds of Google Cloud, Anthos is a cross-cloud/ground platform based on Kubernetes that can be deployed in data centers inside the enterprise. Through the single management interface of Anthos, the required applications can be distributed to the assigned ground terminals or cloud data centers; the service mesh provided by Anthos can even provide cross-cluster collaboration services. In terms of AI, Google Cloud has released dedicated AI tools for different applications, including Kubeflow where training and reasoning all take place on the ground terminal, the Hybrid AI that can perform reasoning on the ground terminal directly without the need for learning, and the AutoML that can learn the reasoning of the ground terminal in the cloud. Through these tools, enterprises can build the best smart manufacturing systems by themselves.
One application of the AutoML of Google Cloud was even exhibited on-site to detect flaws on the production line. The order in which the model training process of machine learning in the past took place was to define the goal, collect and process data, create the model, optimize the model and deploy the model, etc. In the process described above, the last three steps required consuming large amounts of time to repeatedly adjust the parameters, and this is usually the main reason for the delayed launch of the system. Through AutoML, these three processes can be completed automatically. The tool is easy to use; engineers only need to clearly define the learning conditions, they do not need to write programs or have too much professional knowledge of machine learning to quickly complete the training model for their own machine learning model.
Finally, there was the implementation and application of data analysis in smart manufacturing. Google Cloud used the modern data warehousing tool BigQuery for illustration. The amount of data has increased significantly in the smart era, and to handle such large amounts of data properly and apply them accurately are not easy tasks. BigQuery is a data warehousing service that does not require server management; enterprises can use its parallel computing technology to analyze the big data efficiently. During the presentation, a large global IC design company was used as an example; it enhanced the flexibility of its IT system significantly through BigQuery and lowered the system cost.
As the intelligentization trend has become dominant, establishment of the cloud is indispensable for the new generation of manufacturers. For future manufacturing systems, the cooperation between Google Cloud and iKala can not only satisfy the requirements of manufacturers for performance, stability, security and cost, it can also provide various AI tools to help manufacturers build optimized systems successfully and seize the tremendous business opportunities of Industry 4.0.