Company

CHANGSHIN INC is a global footwear manufacturing specialist that provides end-to-end production capabilities across planning, development, and manufacturing. With advanced quality management and flexible production capabilities, the company helps global customers respond to fast-changing market demands while strengthening competitiveness through sustainable growth and continuous innovation.

Challenge

CHANGSHIN INC needed a more scalable way to collect, integrate, and analyze equipment, sensor, and PLC data while identifying potential equipment anomalies in real time instead of depending primarily on periodic manual inspections.

• Need for a framework to collect, analyze, and utilize equipment data

  Although data was being generated from sensors and PLCs attached to equipment, a structured framework was required to collect and analyze this data comprehensively and utilize it for decision-making.

• Equipment management centered on manual inspections

  The existing approach largely depended on periodic on-site patrols and experience-based inspections conducted by field personnel.
  The company aimed to improve equipment management efficiency by detecting signs of anomalies in advance.

• Need for real-time anomaly detection

  A structure was needed to continuously monitor equipment conditions and enable personnel to immediately recognize any signs of anomalies.
  The final report also identified a real-time unmanned monitoring framework and an alarm-based response system as key directions, moving away from conventional periodic on-site patrols.

• Need for vibration data-based equipment diagnosis capabilities

  Vibration sensor data is critical for identifying equipment abnormalities and early signs of failure.
  However, converting vibration data into meaningful diagnostic information required defining diagnostic parameters, failure types, diagnostic rules, and a failure library.

• Need to integrate sensor, PLC, and legacy data

  Accurately assessing equipment conditions is difficult with data from a single sensor alone.
  A structure was required to collect and analyze vibration sensor data, temperature sensor data, PLC data, and existing system data together.

• Need to standardize diagnostic results

  It was necessary to reduce variations in diagnostic results caused by differences in personnel experience.
  A framework was required to standardize failure modes and diagnostic rules and assess equipment anomalies based on consistent criteria.

• Need for a stable data collection architecture considering overseas manufacturing sites

  Data collection and transmission need to remain stable even in overseas factory environments.

Solution

DDI designed and implemented an AWS-based cloud architecture that combines industrial data collection, predictive maintenance analytics, and generative AI diagnostic reporting. The platform enables CHANGSHIN INC to monitor equipment conditions, detect anomaly signals, and support maintenance decisions with data-driven insights.

• Built an AWS-based cloud equipment monitoring platform

  [Data Processing Architecture Based on Key AWS Services]

  

  [Global Network Architecture]

  

• Applied data processing architecture based on key AWS services

  • AWS IoT Core: Serves as the data collection channel connecting equipment and sensor data to the cloud.

  • Amazon S3: Serves as the data repository for collected equipment data.

  • AWS Lambda: Performs data preprocessing, equipment anomaly detection, and equipment anomaly diagnosis.

  • Amazon RDS: Serves as the database for managing diagnostic results and operational data.

  • Amazon EC2: Provides the operating environment for the predictive maintenance solution.

  • AWS Bedrock: Enables AI-based equipment diagnosis and diagnostic report generation.

• Implemented an edge-based equipment data collection framework

• Implemented collection and preprocessing of vibration sensor and PLC data

• Applied DDI’s predictive maintenance solution

• Implemented an AI-based equipment diagnosis process

• Automated diagnostic reports using generative AI

• Standardized diagnostic rules and failure modes

• Customized user interfaces and enhanced business usability, including dashboards, monitoring, search, reports, and diagnostic rule management

• Established a managed service transition and maintenance framework

Following the pilot launch, DDI prepared a phased transition plan for operations and established managed service procedures, including diagnostic report issuance, alarm email delivery, system usage monitoring, error and improvement tracking, and diagnostic rule validation and enhancement.

Benefit

By combining real-time equipment monitoring with AI-based diagnostic automation, CHANGSHIN INC improved maintenance visibility, reduced reliance on manual inspection workflows, and established a scalable foundation for data-driven equipment operations across manufacturing environments.

• Secured real-time visibility into equipment conditions

  Equipment data can now be collected and monitored in real time.
  Field personnel can immediately check equipment conditions through dashboards and monitoring screens.
  When an anomaly occurs, alarms enable personnel to quickly identify the affected equipment and diagnostic status.

• Improved efficiency of manual inspection tasks

  The company can shift from periodic on-site patrols to a data-driven monitoring framework.
  Rather than repeatedly inspecting all equipment in the same manner, personnel can prioritize equipment where signs of anomalies have been detected.

• Reduced expert workload through AI-based diagnosis

  AI analyzes complex vibration waveforms and equipment parameters to generate draft diagnostic reports.
  Instead of drafting reports from scratch, analysts can focus on reviewing and refining the results proposed by AI.
  This reduces the time required for expert intervention and improves diagnostic productivity.

• Enabled faster recognition of and response to equipment anomalies

  When signs of equipment anomalies occur, alarms allow personnel to recognize the situation immediately.
  By reviewing diagnostic reports together with equipment condition data, personnel can respond selectively and effectively.
  Instead of inspecting all equipment uniformly, the company can focus its response on higher-risk equipment.

• Standardized failure types and diagnostic results

  Equipment anomalies can be assessed consistently based on failure modes and diagnostic rules.
  Diagnostic variation caused by differences in personnel experience can be reduced.
  Standardized diagnostic results can be used to establish future maintenance plans and enhance equipment management standards.

• Established a foundation for a maintenance history database

  Diagnostic results and personnel feedback can be accumulated to build a maintenance history database.
  The accumulated data can be used to analyze recurring failures, identify equipment vulnerabilities, and determine maintenance priorities.

• Enabled continuous enhancement of diagnostic rules

  Users can provide feedback on AI diagnostic results.
  This feedback is used to validate and improve the accuracy of diagnostic rules.
  As operations continue, the diagnostic framework can evolve to reflect equipment characteristics and field experience.

• Supported data-driven decision-making for equipment operations

  Equipment conditions, anomaly history, diagnostic results, and maintenance history can be reviewed as data.
  Managers can make equipment operation and maintenance decisions based on data rather than relying solely on experience or intuition.
  This aligns with CHANGSHIN INC’s objective of establishing a decision-support system for equipment operations.

• Secured a foundation for expansion across global manufacturing sites

  An AWS-based system architecture was established with consideration for users in Korea and global locations.
  After validation, the foundation is in place to expand the solution to other equipment or factories.

• Strengthened operational stability and maintenance responsiveness

  Following service launch, a framework was established to support service stabilization and managed services.
  Operating risks can be reduced through response procedures by incident severity, technical support, and emergency support processes.