Building Your Own MES System

Building Your Own MES System with WinCC OA: A Journey of Challenges and Innovations

I've been reflecting on a significant project from my past where we decided to build our own Manufacturing Execution System (MES) using WinCC OA, and I wanted to share the experience—both the triumphs and the challenges.

Why We Built Instead of Bought

Our client manufactures large aerospace components using massive CNC machines—parts up to 12 feet in diameter that can take weeks to produce. Management had minimal visibility into machine statuses, downtime reasons, and operator performance, especially during overnight shifts, making accurate lead time estimates difficult.

While we needed an MES to solve these problems, off-the-shelf solutions weren't suitable because they cater to high-volume production measured in parts per minute, like bottling lines. Our unique challenges included:

• Long Production Cycles: Production was measured in days per part, sometimes weeks.
• Complex Operations: Each part required over 100 distinct CNC steps needing real-time tracking.

Building our own MES allowed us to:

• Provide Real-Time Visibility: Monitor each CNC operation as it's completed.
• Accurately Estimate Lead Times: Offer precise updates based on real-time data.
• Track Operator Performance: Improve oversight, including during overnight shifts.

By creating a tailored solution, we effectively addressed our specific manufacturing needs, leading to greater efficiency and customer satisfaction.

Technical Journey

We developed a distributed WinCC OA system with touch panel HMIs running local instances of WinCC for each of the dozen-plus CNC machines. These were standardized and templated applications. A central WinCC OA server handled data collection, featuring a unified namespace using remote tags in a WinCC cluster. This setup allowed us to aggregate machine data and generate comprehensive reports from clients launched from the central server.

Key components included:

• Kepware Connectivity: We connected to Fanuc CNC controllers to gather current machine states.
• Custom Programs: Developed tools to estimate completion times for each CNC process step, enabling us to calculate performance speeds.
• Unified Namespace: WinCC OA's ability to connect databases, OPC servers, and provide remote tag providers was crucial in unifying namespaces across all remote WinCC edge instances.

Challenges Faced

• Steep Learning Curve: WinCC OA is quite verbose with a high learning curve. I invested time in training to get up to speed.
• Complex SQL Queries: Some SQL queries took days to write, involving intricate self-joins. Stack Overflow became my best friend during this phase.
• Resource Constraints: With a tight budget of around $100,000—including hardware, software, integration testing, and development—we had to be strategic. This budget was for a pilot covering just four machines.
• Technical Hurdles: We created stored procedures and internal API script files to normalize and modularize all code for MES functions like addMachine, getMachineStatus, and addProductionEvent.

Solutions and Learnings

• Modular Coding: By identifying where unique code belonged versus modular code, we templated our applications and scripts effectively.
• Unified Namespace (UNS): Although I didn't know the term at the time, implementing a UNS was pivotal in organizing our system architecture.
• Deep Dive into OEE: Understanding Overall Equipment Effectiveness (OEE) deeply and calculating it accurately was essential. Digital OEE calculation is a must—manual methods often lead to inaccuracies.
• Leveraging Community Resources: Stack Overflow and industry forums were invaluable for troubleshooting and learning best practices.

Outcomes

While I can't share specific metrics, I can say that we started capturing common downtime reasons and provided analytical screens for operators and management. Operators began improving performance simply by visualizing their current production run status in real-time. The transparency and data accessibility made a significant difference.

Moreover, the company I worked for transformed this project into a product offering for other organizations facing similar challenges.

Advice for Others

If you're considering building your own MES system:

• Invest in Learning: Understanding the tools and concepts deeply will save you time and headaches down the line.
• Modular Design is Key: Templatize where possible to streamline development and future updates.
• Digital OEE is Non-Negotiable: Accurate, real-time calculations are crucial for meaningful insights.
• Join a Structured Program: Reflecting on this journey, I realize how beneficial it would have been to have guidance.

That's why I'm excited to announce that I'll be guiding an 8-Week MES Bootcamp Accelerator, alongside Walker Reynolds, who has also developed an MES 4.0 application. This program compresses what could be a 6-month learning curve into just 8 weeks, saving you time, effort, and resources while helping you avoid common pitfalls.

Why Build Over Buy?

In hindsight, building was the only viable option for us. Our unique processes and the scale of our operations meant that off-the-shelf MES solutions weren't adequate—especially when production counts were measured not in parts per minute, but in days per single part.

If you're grappling with similar challenges or considering building your own MES system, I invite you to join us in the upcoming bootcamp. Let's navigate this complex but rewarding journey together.

Signup for our MES Bootcamp Free webinar here: https://www.iiot.university/mba-optin

Feel free to reach out if you have any questions or would like more information!

Zack Scriven | Digital Transformation Coach at 4.0 Solutions