A Borg company

Solved: PLC with High Speed, High Volume Data Tracking

How do we generate, track and add value to process and product data at very high speeds in a PLC?
Solved: PLC with High Speed, High Volume Data Tracking

Some time ago, we were approached by a customer looking for a custom high-speed solution that would assist them in tracking product. The request and process were of such a nature that we had to sit down and ask ourselves the following question.

How do we generate, track and add value to process and product data at very high speeds in a PLC?

Since the advent of the PLC in the late 1960s process industries and manufacturing have been rapid adopters for this technological advance over relay logic. Hand in glove with industrial automation and the microchip has been the massive growth of the internet and the personal computer. Gone are the days of one computer being used by scientists in isolated laboratories. Today companies generate product and knowledge, but more than ever they generate data.

Consider a bottling line. If each bottle has a barcode that identifies product type, batch and date, we might need to track those details at a rate of 1,000s of bottles per batch. Not a difficult task, we just receive a set of details for that batch when the machine or work center starts that product run. What if we choose to uniquely identify a single bottle? Sure, we could have a starting number, increase that number by one for each time the bottle counter indexes and we are done right?

Not so much.

What happens if your process removes a bottle that is under-weight for example? Do we re-print a label and have two printed labels with the same identifier on them? What happens if that QA failed bottle ever made it out of the factory?

What would be perfect is tracking a line of bottles as a conceptual model of data, with a record for every position on the bottle line.

At this point, the more IT focused controls specialist starts to contemplate questions like ‘Why track that data in the PLC, wouldn’t it be easier in the database or software world?’

Yes, it would be, if speed weren’t an issue.

But how do you scale this further? Say you now have 10 lanes of bottle production and you want to have the computer vision results on the label, including quality and readability of the expiry date digits and date as a whole string moving around your system? All of this needs to happen while the line is moving at 2m/second.

The MES must provide the unique numbers for your bottles and receive back data from points along your production line, such information as QA data, vision inspection system data, label placement, content and accuracy. Again all this information needs to be tracked and stored. To complicate matters further, consider a scenario where bottles may be considered as waste/defect half-way through your process, but you still need to track all the data and have the bottle separated at a point further down in the chain for manual inspection, while all your data must still be returned to the MES for record keeping.  All these steps and processes must occur at speeds and volumes that make pre-caching the data before a run, and saving it after the run completes, completely unfeasible.

At Crossmuller we have been fortunate to be involved in developing and implementing a system such as  this.  A system where the solution needs to manage thousands of records, each unique and tracked from a generation of serial numbers for pre-approval, to closing the QA loop with images of faulty printing on labels. Managing rates of 200 products a second for sustained throughput while integrating multiple inspection points, process work units and vendor machines.

A project of this nature starts with a few critical steps:

  • Analysing  your process in detail. You as the customer are subject matter experts in your process, and we need to gather all the detailed information from your engineering and operations staff to make a fit for purpose design for your system.

  • PLC hardware should be selected to meet your specific requirements. In the case outlined, a Beckhoff PLC was used. Extensive testing and performance profiling of several techniques yielded results using Beckhoff TwinCAT3 PLC hardware that could not be matched at a similar price-point by other vendors.

  • Identifying the correct operational systems. An optimised Wonderware MES implementation and Microsoft SQL Server database was used to perform the manufacturing execution and data management requirements.

  • Identifying vision systems that can handle the speed. High speed line scanning cameras on dedicated FPGA based processing boards were used for image processing. Field data, images, OCR, Quality and machine information were encoded in the data and processed at real-time speeds for sustained periods.

Following the above steps, resulted in a system which can track and report on a product run of tens of thousands of products with incredible detail, at speeds which PLC based control systems traditionally would be pushed to the limit just controlling.

We are proud to have proven that such a system is possible and reliable. Moreover, we found that using this combination of hardware and software, a user friendly, high performance and maintainable system could be implemented at a surprisingly low price.

Crossmuller has over 20 years of experience working closely with Australian manufacturing businesses across multiple industries.  If you’d like to speak to someone about how to start your digital journey, please contact us – we’d love to share some of our experience. 

More Articles

MES: It is also about the journey
MES: It is also about the journey
The MES journey should always start with knowledge of what you have, and where you want to go.
The Benefits of Finite State Machines in Industrial Automation
The Benefits of Finite State Machines in Industrial Automation
Developing a programmable logic controller (PLC) program from scratch can be a long and daunting task. By using finite state machines (FSM), not only is the programming simplified, but it creates a fl...
Photogrammetry Modelling
Photogrammetry Modelling
Photogrammetry Modelling - Visualising Tomorrow using Digital Technologies