Smart Factories and Automation

Smart Factories

To enhance factory efficiency, improve manufacturing process quality and meet customer delivery time demands, ASEH began to invest in automated, lights-out factories in 2015. Automation, heterogeneous integration in machine and production systems, and heterogeneous integration in systems-in-package (SiP) are 3 major forces driving smart factories and digital transformation at ASEH. In 2011, ASE established the ASE CIM Committee comprising of automation 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 Management Center. By 2021, the company has 25 lights-out factories in operation, with more than 600 automation engineers trained, and over 45 industry-academia research projects developed.

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	At the time when the ASE CIM Committee was established, there were only 38 engineers who qualified to manage the automation process.	More than 600 smart factory automation engineers have been trained through the establishment of automation and AI academies as well as industry-academia research programs.

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) (Predictive Maintenance, 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 widespread application of AI

In 2018, ASE 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.

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.

Automation technologies introduced in 2021

Introduce the popularization of AI platform

The IAI (Industrial AI) platform hopes to quickly introduce AI learning, pass on experience through AI, and create a company culture of national AI. As long as data is provided, different prediction models can be built to achieve the purpose of production capacity, quality, and process prediction.

RPA (Robotic Process Automation)

RPA is a software bot uses a combination of automation, computer vision and machine learning to carry out a ton of mundane and repetitive computer-based tasks at error-free, high volume and speed that are previously handled by humans.

RTD (Real Time Dispatcher)

RTD (Real Time Dispatcher) aims at maximizing efficiency and output, considering various production conditions, using powerful computers and big data to develop intelligent dispatching systems and automatic dispatching systems to shorten the production cycle.