Challenges of Taiwan's Manufacturing Transformation to Smart Manufacturing

In the face of rapid and iterative new technologies, how can an enterprise stand firm in the market without being eliminated, invest, and consolidate its wisdom to create a new future? Taiwan ’s general manager of IBM said that a few months ago IBM invited the CEOs of 100 leading companies worldwide to discuss the future development of the betting focus. She suggested that companies should not blindly follow new technologies, re-examine their core competitiveness, and then decide which technologies to use for development Smart manufacturing. Two years ago, the company introduced Industry 4.0, started smart manufacturing, and started to make results from a single application on the production line, such as robotic arms or computer vision inspection. However, it was found difficult to copy to other production lines and scenes.

Four Challenges for Taiwanese Enterprises to Develop Smart Manufacturing

A partner of the IBM Global Enterprise Consulting Service Group in Taiwan said: "The dilemma of global smart manufacturing is that nearly 70% of them are difficult to promote and scale quickly." The lack of vertical integration and horizontal expansion makes it difficult to produce specific economic benefits, making Smart manufacturing has encountered a bottleneck, and further pointed out that in the process of developing smart manufacturing in Taiwan, the manufacturing industry generally faces the following four major challenges:

• Challenge 1: Automation should not be the only positive solution to the old factory system.
  The manufacturing industry has a lot of layout automation, hoping to replace manpower, increase the yield of products, and set off a wave of unmanned factories and light-off factories. "Smart manufacturing requires a complete vertical field to produce a certain economic scale and vertical efficiency, which will be the key to successful transformation!" If only automation is developed, it is not structured and scenarios with the supply chain, production planning, material planning, and other processes Integration, such as equipment upgrades, creating data interfaces, etc., will face the issue of whether automated capital expenditures meet the return on investment. Li Liren suggested that, in response to the current status of the enterprise and the market to be developed, the future overall structure and the integration of new technology should be drawn.
• Challenge 2: The economic scale and benefits of AI
  Smart manufacturing have developed a mature single scene, such as machine vision used in yield improvement, defect detection, predictive maintenance, etc., but why can't it still improve the company's yield and save manpower? It is because the single-point results have not formed an economic scale and can be copied to other production lines, or lack of high economic efficiency applications and combinations. Li Liren suggested that enterprises use investment return rates and enterprise KPI to guide value verification and structural integration to accelerate the implementation.
• Challenge 3: The burden of talents and huge old systems.
  For smart manufacturing to develop vertical integration, it will face IT architecture and more than 80% of application systems are issues of old systems. Except that IT budgets are placed on maintaining machine systems, even internal talent Skills also focus on the old system. It is necessary to make good use of ecosystem partners, execute quickly, and expand rapidly. Once it is time to move on to the development of digital transformation and smart manufacturing, it is necessary to train new skills for internal talents. If the economic scale of the group is large enough, you can invest in a new project team with a clear role of integrator, and use newly hired talents for rapid verification, deployment, and promotion, so that the scene application can be quickly achieved. If Dong always manages to take the initiative, he can accelerate the benefits.
• Challenge 4: Vertical integration and horizontal diffusion.
  When companies develop smart manufacturing, they often start with project trials and cut application scenarios very fragmented, while the initial investment and results have become burdens for subsequent development. Besides, without the future integration structure, it is difficult to quickly spread to different production lines or group departments. The enterprise should establish a clear and complete execution blueprint and timetable, and the execution plan should take three years to plan the execution budget every year.

A Japanese industrial robot manufacturer wanted to develop a new business model and decided to invest in smart manufacturing, build smart factories and robot automation, and quickly expand to the market. This Japanese industrial robot manufacturer will integrate strong OT (Operational Technology) into IT to make a complete vertical integration structure, quickly step out of a single factory, and expand rapidly. To develop a digital factory, China National Heavy Machinery Group plans a four-step overall blueprint.

• The first step is to do a status quo assessment, complete inventory of existing enterprise resources, and analyze the structure of the digital factory.
• The second step is a business improvement and demand analysis, put forward management improvement suggestions and objectives, identify the key system support points and needs of the digital factory, and determine the enterprise's digital factory business model.
• The third step is the overall digital factory planning architecture, including a complete digital factory planning blueprint, strategic goals, and establishing a digital factory application architecture that matches the business, as well as related basic technologies and systems.
• The fourth step is the digital factory, determine the project to be achieved, and list the detailed implementation plan, focus on the goals in a three-year rolling and one-year adjustment, and establish a perfect digital factory management system to create a system supervision system to the system guarantees success.

5C maturity model: Review the process of enterprise smart manufacturing

How does the manufacturing industry examine the depth and breadth of the company's smart manufacturing process? IBM launched the "5C maturity model" for smart manufacturing, which is divided into five stages by maturity:

Step 1: Device connection (Connect): use the Internet of Things and machine networking to achieve device-to-device connection and collaboration. It is difficult to develop AI and application scenarios if the old equipment is not smart enough to collect data.

Step 2: Data conversion (Convert): combining AI big data platform and edge computing, to develop intelligent AI application scenarios on the device side. For example, AI visual inspection, AI predictive maintenance.

Step 3: Predictive simulation (Cyber): Introducing digital twins (Digital Twin) to the digital factory. Visualize the production site, and dynamically simulate to scheduling, and even achieve dynamic scheduling of learning engine orders and production lines.

Step 4: Smart Factory (Cognitive): To build an artificial intelligence learning platform, so that the factory can self-diagnose, self-repair, automatic scheduling, and accelerate model verification and deployment.

Step 5: Dynamic customization (Configure): through B2B hybrid cloud platform, blockchain, dynamic customization, to achieve a small variety of short delivery time, to achieve a software-defined value chain platform.

As the scenario application becomes more mature, enterprises face four major challenges when investing in smart manufacturing. The biggest bottleneck is a vertical and cross-product line, cross-sector integration. IBM recommends that enterprises plan the overall blueprint and match with suitable partners at each stage Accelerate the implementation of smart manufacturing.