AWS for Industries

Designing an integrated production monitoring and analytics platform to improve Jobs Per Hour and rework ratio

Problem statement

A leading OEM (Original Equipment Manufacturer) faced a persistent challenge of low JPH and high rework ratio in the vehicle production process, spanning the inbound stage to final inspection. JPH (Jobs Per Hour), is a metric commonly used in automotive production, to measure efficiency of a production line. The current JPH (Baseline) was 10. The aspiration was to raise that number to 16 by the end of the two quarters following the rollout of a structured manufacturing excellence program. The previous rework ratio was around 6 percent while the goal was to reduce that to 2 percent by the end of the subsequent two quarters’ production. The core engagement team comprised of LoB (Line of Business Leaders like quality and production heads), general manager IT along with AWS and AWS partner team.

During the discovery phase, a team of domain specialist from Amazon Web Services (AWS) worked with the customer and identified the following operational gaps leading to high rework ratios and not meeting takt time: 1) siloed data across processes, hindering effective analytics on JPH losses; 2) a lack of short interval controls for implementing countermeasures against production shortfalls, 3) a legacy platform that could not perform root-cause analysis of losses at a given point in time; 4) imbalanced production capacities across assembly, body shop, and paint shop, leading to intermediary inventory pileup; 5) production supervisors spending, on an average, two working days each week on data entry and eight working days each month doing rework reasons analysis with manual spreadsheets; and 6) an inadequate process for entering delays from the production line and quality inspection station into a single database of stoppages, leading to variations in takt time reporting and failure to integrate and resolve andons.

Proposed solution

AWS proposed deploying a unified, integrated solution for production monitoring and analytics across the assembly, body shop, and paint shop operations. As shown in figure 1, the proposed platform had three core parts: 1) hardware integrations; 2) a core application; and 3) a centralized control tower.

Figure 1. Core parts of the proposed solution for integrated production monitoring and analyticsFigure 1. Core parts of the proposed solution for integrated production monitoring and analytics

Part 1: Hardware integrations for the proposed platform

Integration of the shop floor PLC (programmable logic controller) systems with an edge PLC for taking process parameters, on both on-demand and update-interval bases using OPC UA and OPC DA as protocols. Also required were modifications in the existing PLCs, which were provisioned according to present license modalities. Additional hardware components included the following:

  • Network cable/tray laying and connecting to an edge server along with commissioning of the new PLC system to access the process parameter which were earlier not captured;
  • A barcode reader, along with SAP terminals, provisioned at specific stations to automate data capturing, thus reducing manual effort and entry errors. We’ve called this new system the Operator Assistance Module (OAM);
  • Touchscreen terminals/mobile apps to interact with internet of things (IoT) application screens for data entry/viewing; and
  • A QR code pasted onto each repair bay to track rework quantity, rework time, and rework consumables.

Part 2: The core application alongside the hardware elements of the proposed platform would enable the following to track the JPH losses:

  • Starting from the body shop and general assembly, the application powers an operator assistance system that allows operators to select errors or issues from a predefined list. That is then complemented by defects-and-errors tracking and visualization, indexed by the VIN (vehicle identification number), to drive short interval controls for line managers.
  • The application also facilitates a detailed root-cause analysis of errors and defects on the line by triggering a CAPA (Corrective Action and Preventive Action) workflow immediately upon acknowledgment from operators through a digital workflow system called Digital Andon cards.
  • Utilize existing data lake for operational technology data and enterprise application data consolidation.
  • IoT events hub to manage IoT sensor data from general assembly, body shop and paint shop assets, paint, and inspection processes data, 3rd party sensor data to help identify anomalies and deviations in real-time. By identifying and reducing these anomalies the overall loss time impacting JPH can be further minimized.
  • Streaming analysis for low-latency and near real-time analytics of inspection and assembly operators behavior and respond almost instantaneously with relevant recommendations thus eliminating the root casus for rework upfront.
  • A list of existing potential failure modes is visualized on the screen to aid operators in considering countermeasures.
  • A list of performance metrics is presented on a dashboard to aid decision-making by shift supervisors and production managers. These metrics include, for example, linewise JPH on a near real-time basis, takt time, deviation analysis, mean time between failure and repair, bottleneck analysis from a capacity standpoint, and more.
  • Self-service historical data analysis is available, showing statistics, correlations, and trends of KPIs (key performance indicators) with respect to various dimensions and attributes of the vehicle. These attributes if not in line with the specifications can be course corrected in real-time thus reducing the extent of rework ratio.
  • An OAM provides a relevant standard operating procedure (SOP) in the form of a workflow (or discrete instructions) based on a user-inputted query or the current situation on the floor to aid the operator in taking action during proactive inspections. This impacts the rework time per vehicle produced.
  • A digital check-sheet for machine start-up and changeover is available, helping reduce the time needed to manually handle information and bringing it closer to error-proof. This reduces the unplanned management losses thus impacting the JPH.
  • In addition to AWS Connector for SAP, the application leverages best practices and top tools for communication encryption (TLS 1.2), data encryption, and authorization management with AWS Identity and Access Management (AWS IAM)—which other businesses use to securely manage identities and access to AWS services and resources.
  • Using metrics for available/running capacity and orders at hand, an order-planning tool generates production schedules on a shift-wise, daily, or weekly basis.

Figure 2. A table showing categories for KPIsFigure 2. A table showing categories for KPIs under the application’s scope across all three subsections of the manufacturing setup to aid decision-making by the CEO, the COO, and the manufacturing team

Part 3: A virtual “Andon” system to help drive short interval control

The application powers an integrated virtual Andon system that exposes APIs so events and issues can be raised by operators around JPH loss time and rework reasons. These issues can then be routed using a work flow in real-time to the resolution team or individual. Additionally, enhancements are planned to help improve the user experience including ability to add images for easy recognition of issues, ability to add detailed messages for root causes for quicker guidance on resolving similar issues and ability to view and track issues across a manufacturing site or specific location or areas. Administrators (such as supervisors and production managers) use a web interface to define their factory setup, site name, process type, event types for each process, and lists of workstations. In the solution’s workflow, users can monitor manufacturing workstations for an event, log the event, and then route the event to the correct engineer for resolution.

Conclusion

The solution outlined in this post has a potential impact of up to 24 percent on overall JPH across all vehicle types. That, in turn, has a direct impact on the daily cost of production to the tune of a 1.1 percent and 15 percent impact on the time spent on rework. The percentages quoted were based on the estimations made along with the OEM’s industrial engineering team over and above the past 12 months normalized base lines analyzed. The project also identified bottlenecks in capacity planning and scheduling at the shift and daily levels. The simple scheduler application helped production managers plan according to incoming raw materials, inventory at hand, intermediate inventory, and available manpower and generated schedules according to worker skill sets.

Similar case studies and success stories, spanning the manufacturing and supply chain vertical, can be found at https://aws.amazon.com/automotive/. In addition, you may reach out to us for an assessment of your current JPH and rework baselines and co-develop a solutions roadmap.

Appendix

Definitions

  • Takt time—a fundamental concept in lean manufacturing; a measure of the maximum amount of time that can be taken to produce a product while still meeting customer demand.
  • Andon—a visual and auditory alert mechanism used in manufacturing environments; signals the status of the production process and indicates when immediate attention is required (source: Lean Tools).
  • CAPA—a methodology for managing and resolving quality issues in products and services; involves corrective action to address existing problems and preventive action to avoid future risks (source: Lean Six Sigma).
  • Rework—the outcome of poor quality; assessed according to in-line defects and errors reported during inspection of a vehicle.
  • Production performance—actual JPH/target JPH
  • General Assy – General Assembly is a section for vehicle assembly with in the production set up.
Anindya Bhattacharya

Anindya Bhattacharya

Anindya Bhattacharya, Industry Specialist Manufacturing and Supply Chain at AWS, drives the Manufacturing and Supply chain Industry solutions for the AWS India market. He brings in close to two decades of core Manufacturing and Supply chain expertise focusing on delivering holistic EBITDA transformation, Smart Manufacturing deployments and advisory around strategic Supply chain platforms. Prior to joining AWS he has worked with Hitachi, Blue Yonder, EY and TATA STEEL group in various capacities globally. Anindya has hands on experience of delivering large scale and piece meal Smart factory deployments across key industry segments like Metals, Automotive and Industrials segment.

Ayush Ragesh

Ayush Ragesh

Ayush Ragesh is a Solutions Architect at Amazon Web Services with over 9 years of overall cloud experience, including 3.5 years helping automotive and manufacturing customers adopt cloud technologies. His specialization in IoT technologies, combined with his expertise across various domains, allows him to support these customers in driving innovation and embracing cloud adoption. Ayush's passion for cutting-edge technologies has always enabled him to be a quick learner, staying ahead of the curve in an ever-evolving landscape.