Smart Factories and Automation
Illuminating Excellence in Smart Manufacturing
Aiming to drive greater efficiency and improvements in our manufacturing process that will in turn deliver higher customer satisfaction in quality and delivery, ASEH began to invest in automated, lights-out factories in 2015. At ASEH, we are accelerating digital transformation in smart manufacturing through automation, heterogeneous integration in machine and production systems, and heterogeneous integration in systems-in-package (SiP). In 2011, ASE established the ASE CIM Committee, a strategic task force that is comprised of teams from various business units (lead frame packaging, ball-grid array packaging, flip chip packaging, wafer-level packaging, SiP packaging and test services) and the Information Technology Center. By 2023, the company has established 46 lights-out factories, trained more than 700 automation engineers, and developed over 57 industry-academia research projects. ASEH achieved another major milestone when its bumping facility in Kaohsiung was inducted into the World Economic Forum's Global Lighthouse Network, a community of production sites and value chains that are world leaders in the adoption and integration of the cutting-edge technologies of the Fourth Industrial Revolution (4IR).
White Paper on Packaging and Testing Machine Safety in a Smart and Sustainable Factory
As a newly inducted member of the globally recognized WEF Global Lighthouse Network (GLN), ASE Kaohsiung is deeply committed to continuing its dual-axis transformation that integrates smart manufacturing and sustainable development. This is achieved by seamlessly incorporating cutting-edge AI and other advanced Industry 4.0 technologies into daily operations, harmonizing production optimization with environmental sustainability. In the realm of smart manufacturing, smart scheduling is implemented to replace manual labor, significantly improving efficiency and reducing product manufacturing cycles. AI image recognition, machine learning, large language models, and feature engineering technology are utilized in process control and management to generate early warnings on equipment anomalies and predicting equipment lifespan, ensuring peak yield rates, reduced processing time, and risk mitigation. In sustainable development, AI-enhanced smart energy management is implemented to efficiently adjust manufacturing equipment according to environmental conditions and production requirements. In addition, we employ collaborative AI computations to optimize water operations and increase water recycling, thereby reducing water consumption. Through partnerships, we optimize waste resource utilization in alignment with the circular economy model, ensuring appropriate waste removal without environment impacts by using AI tools to monitor waste transport vehicle operations. As a people-centric organization, ASEH prioritizes a safe, secure, and healthy workplace environment. In collaboration with our facilities, subsidiaries, and 6 industry peers, we spearheaded the inaugural White Paper on Packaging and Testing Machine Safety in 2023, targeting the semiconductor packaging and test sector. To identify safety risks in machinery and design, the white paper delved into domestic and international safety standards and collaborated with government, industry, and academic stakeholders to identify occupational hazards and propose preventative measures. The paper addresses human, machine and environmental factors and focuses on safer design and ergonomic hazard prevention to enable a culture of workplace safety and employee well-being.
Innovative and Breakthrough Methods Adopted in the Creation of Smart Factories
Challenge
Problems encountered
Solution
Inadequate equipment connectivity
To meet the needs of smart factories, production equipment information must be collected and stored in a central database so that real-time analyses and management can be conducted.
In the early days, due to the dearth of OSAT industry production equipment that met Semiconductor Equipment Communication Standards (SECS), equipment connectivity was the top challenge to be overcome.
Step 1
Collaborate with procurement units to conduct negotiations with equipment suppliers and request that new production equipment meet SECS standards.
Step 2
Perform research on existing production equipment to find ways to achieve automatic connection and convert into compatible SECS formats. After years of development, ASEH’s production equipment now meets SECS standards.
High complexity of product tracking
Automotive customers require strict records of the production history of all automotive chips to facilitate tracking when problems occur.
In semiconductor chip manufacturing, product tracking begins at the wafer fabrication stage. The wafers will then proceed to the next process stage. Once the wafer is cut into individual dies for packaging, the dies do not have any markings for identification and tracking.
Use 2D codes and RFID technology to accurately record the individual wafer and the location on the wafer that each die originated from, the location on the substrate and the locations on the die carrier and substrates.
All the location information are stored in the map system database that can be accessed any time. Customers are able to check production history, while our engineering teams can use the data to perform quality and yield analyses.
Lack of local automated equipment supply chains
In the early stages, most automated equipment suppliers were large foreign suppliers that commanded high prices, were inflexible and provided long lead times. As a result, we faced delays in project completion and unsatisfactory outcomes.
Actively look for local suppliers of automated equipment including automated guided vehicles, automatic storage and robotic arms, etc. In recent years, we have established business relationships with approximately 38 automation suppliers, strengthening the local automation industry chain in Taiwan.
Lack of qualified personnel
When the ASE CIM Committee was initially established, there were only 30 engineers qualified to manage the automation process.
More than 700 smart factory automation engineers have been trained through the establishment of inhouse automation and AI training modules as well as industry-academia research programs.
AI training modules. We launched the modules in 2018 to promote AI technology. Integrating the AI platforms into the production, engineering, and administrative departments help to popularize the IAI platform, and also ensure readiness for the upcoming No-code AI age. As of 2023, more than 10,000 individuals have been trained.
Intelligent Engineering training modules. Since the launch in 2022, our PE/EEs have received training in statistical analysis and equipment monitoring. The engineers also learnt how to optimize digital tools and ideas for project applications. As of 2023, more than 3,000 individuals have been trained.
Digital Application training modules: Starting in 2018, we created courses on digital tools such as RPA, Qlick View, Doc.Bee, and Co-know to train administrative and support staff to utilize digital tools effectively. As of 2023, more than 4,000 individuals have been trained.
Challenge
Problems encountered
Solution
Inadequate equipment connectivity
To meet the needs of smart factories, production equipment information must be collected and stored in a central database so that real-time analyses and management can be conducted.
In the early days, due to the dearth of OSAT industry production equipment that met Semiconductor Equipment Communication Standards (SECS), equipment connectivity was the top challenge to be overcome.
Step 1
Collaborate with procurement units to conduct negotiations with equipment suppliers and request that new production equipment meet SECS standards.
Step 2
Perform research on existing production equipment to find ways to achieve automatic connection and convert into compatible SECS formats. After years of development, ASEH’s production equipment now meets SECS standards.
Inadequate equipment connectivity
Automotive customers require strict records of the production history of all automotive chips to facilitate tracking when problems occur.
In semiconductor chip manufacturing, product tracking begins at the wafer fabrication stage. The wafers will then proceed to the next process stage. Once the wafer is cut into individual dies for packaging, the dies do not have any markings for identification and tracking.
Use 2D codes and RFID technology to accurately record the individual wafer and the location on the wafer that each die originated from, the location on the substrate and the locations on the die carrier and substrates.
Every location data is stored in the map system database that can be accessed any time. Customers are able to check production history, while our engineering teams can use the data to perform quality and yield analyses
Inadequate equipment connectivity
In the early stages, most automated equipment suppliers were large foreign suppliers that commanded high prices, were inflexible and provided long lead times. As a result, we faced delays in project completion and unsatisfactory outcomes.
Actively look for local suppliers of automated equipment including automated guided vehicles, automatic storage and robotic arms, etc. In recent years, we have established business relationships with approximately 38 automation suppliers, strengthening the local automation industry chain in Taiwan.
Lack of qualified personnel
When the ASE CIM Committee was initially established, there were only 30 engineers qualified to manage the automation process.
More than 700 smart factory automation engineers have been trained through the establishment of in-house automation and AI training modules as well as industry-academia research programs.
AI training modules. We launched the modules in 2018 to promote AI technology. Integrating the AI platforms into the production, engineering, and administrative departments help to popularize the IAI platform, and also ensure readiness for the upcoming No-code AI age. As of 2023, more than 10,000 individuals have been trained.
Intelligent Engineering training modules. Since the launch in 2022, our PE/EEs have received training in statistical analysis and equipment monitoring. The engineers also learnt how to optimize digital tools and ideas for project applications. As of 2023, more than 3,000 individuals have been trained.
Digital Application training modules: Starting in 2018, we created courses on digital tools such as RPA, Qlick View, Doc.Bee, and Co-know to train administrative and support staff to utilize digital tools effectively. As of 2023, more than 4,000 individuals have been trained.
Smart Factory Milestones
2011
Introduced the recipe management system (RMS)
As a control measure before mass production, the EAP transfers data to equipment through SECS/GEM, ensuring data validity and improving overall equipment efficiency (OEE).
2013
Inhouse-developed Semiconductor Equipment Communications Standard (SECS) equipment automation program (EAP)
To overcome challenges in equipment connection program development, we designed a development platform for standardized equipment connection programs, solving process design problems, lowering program development complexity, and increasing human-machine ratios and operation time.
Implemented the fault detection and classification (FDC) system
By collecting equipment production parameters in real-time, systems are able to report equipment status immediately and check formal functions automatically so that warning signals are issued when malfunctions occur, thereby preventing the repeated manufacturing of defective products and ensuring that reporting mechanisms are in place to detect malfunctions in real time.
2015
Introduced robotic arms and automated guided vehicles (AGVs)
AGVs and robotic arms were integrated to introduce the autonomous mobile robot (AMR) that can support transport operations, thus reducing manpower on the floor and maximizing packaging capacity.
2018
Ushering in the era of AI
Applying AI powered detection technology to identify and intercept any malfunctioning equipment that may compromise information security and prevent any information security incidents. The inhouse-developed technology helps mitigate information security risks and reduce investment costs.
2019
Incorporated the predictive maintenance system (PdM)
A predictive maintenance system helps determine equipment that is likely to require maintenance and predicts equipment component failures and malfunctions in advance. The system allows early notification of maintenance personnel to service the equipment, thereby lowering equipment failure time.
2020
Launched the world’s first 5G mmWave smart factory
The 5G mmWave smart factory was a collaborative effort between ASE, Chunghwa Telecom and Qualcomm, showcasing the future of automation and smart factories. 3 use cases were developed to demonstrate the use of 5G mmWave in smart factories - automated production line inspection using AI+AGV, remote AR maintenance and the AR experience at the ASE green technology center.
2021
Build an IAI platform to promote the universal application of AI
ASEH ushered in the era of AI. In addition to actively cultivating AI technology talent, we began to build the IAI platform to create an AI no code environment and promote widespread application of AI throughout the company.
2022
Inducted into the World Economic Forum's Global Lighthouse Network
ASE's Bumping Factory in Kaohsiung adopts 4IR (Fourth Industrial Revolution) technologies across its manufacturing operations. In particular, the facility applies AI technology in the management of equipment and processes to improve yield and accuracies in production schedules. As a result of the remarkable integration of 4IR, the facility was inducted into the World Economic Forum’s Global Lighthouse Network (GLN).
2023
Generative AI Adoption for the optimization of manufacturing processes
ASE Kaohsiung’s smart manufacturing is continuously evolving and the team is actively harnessing AI to optimize production processes. Our manufacturing processes for a diverse range of products are complex, and AI adoption is helping to improve worker productivity that minimizes work-in-progress costs and maximizes yield. Applying AI enables better optimization of machine and shipment scheduling to meet delivery deadlines, ensuring the most efficient production schedules in the shortest possible time. The extensive data mining and analysis, combined with the factory's 24/7 operations have resulted in widespread AI applications at ASE Kaohsiung.
A Dual-axis Approach in Smart Manufacturing Transformation and Sustainable Development
ASEH is proud that its Kaohsiung bumping facility was inducted into the WEF Global Lighthouse Network (GLN), the gold standard for AI in manufacturing. We remain committed to a dual-axis approach that focuses on advancements in smart manufacturing and sustainable development simultaneously. The integration of Industry 4.0 technologies with AI will drive greater efficiencies into our operations and accelerate sustainability improvements at scale.
To address the production scheduling of varied products in a smart manufacturing environment, key factors are collected and AI algorithms are used to train and create an optimized production model. The model removes the need for manual scheduling and machine programming; saving time, and maximizing production schedules, productivity and manufacturing efficiency. Advanced process control systems utilizing AI image recognition, machine learning, large language models, and feature engineering enable real-time monitoring of various production information, and trigger alerts in the event of any anomalies. In addition, the system is capable of predicting equipment lifespan, thereby allowing timely maintenance to reduce downtime and impact, and maintain optimal yield.
In line with our sustainability goals, we have implemented smart energy management across our manufacturing facility to reduce energy consumption and greenhouse gas emissions. By applying a comprehensive AI-augmented database and monitoring system to establish automatic control, we were able to efficiently adjust manufacturing equipment such as chillers and filter fans in the clean room according to environmental conditions and production needs. In regard to water resource management, we monitor post-process water discharge with the latest fluorescence identification technology to detect trace organics. Combined with the power of AI computation, we were able to optimize the use of water resources to reduce tap water consumption, prevent wastage, and improve recycling. For our waste management, we work closely with multiple partners to create circular solutions for resource optimization. Digital tracking and AI tools are used to closely monitor the movement of vehicles transporting residual waste, ensuring proper treatment of waste disposal and mitigating environmental impact.
In light of fierce competition and climate change, ASE Kaohsiung is capitalizing on its people-centric culture to advance smart transformation and sustainable development simultaneously. The company is investing in resources to train and upskill its employees, reshaping their value and demonstrating our commitment to the environment and net zero. Most importantly, we hope to lead and influence the industry toward a more sustainable future.
Sustainable Impact of Smart Factories
Our smart factory concept began with a strong foundation in automation, and the heterogeneous integration of customers, suppliers and production processes, to drive the semiconductor industry onto a higher value chain and accelerate technology advancements. Smart factories represent the next leap for the semiconductor packaging and test industry to play an enabling role beyond More than Moore.
Other topics
Other topics
