You need to go deeper than MES - is your data model ready?

You need to go deeper than MES - is your data model ready?

By Mathew Daniel, VP-Operations, Sciemetric Instruments

Mathew Daniel, VP-Operations, Sciemetric Instruments

The key to achieving the performance and quality benchmarks that define Industry 4.0 is giving people on the front line the insight to take action when it needs to be taken.

Workers on the production line need the tools to catch errors, flaws and anomalies as they occur, at any station. This data must be visualized for easy interpretation by people other than a quality engineer, so they can understand the issue and take swift corrective action.

Having the means to catch and address problems where and when they occur is crucial to continuous process improvement. It cuts scrap and rework, and the risk of costly warranty repairs or mass recalls that can tarnish a brand’s reputation.

But to achieve this, key stations on the line must be equipped to collect, analyze and visualize data at a granular level beyond the capabilities of a conventional manufacturing execution system (MES) tied into an enterprise resource planning (ERP) suite.

"Supplementing your MES investment with a robust MP&QM system is well worth it"

According to Frost and Sullivan, this deeper layer of intelligence is called Manufacturing Performance and Quality Management (MP&QM).

Capturing the ‘DNA’ of a process

Typical MES systems rely on SPC (statistical process control) and gather only scalar data. These are isolated data points from each process on a production line such as max torque or final spindle angle. Testing stations may be just binary propositions - green light, pass, red light, fail.

We talk to manufacturers all the time that struggle with this superficial level of data collection to improve quality and first-time yield. They need to collect more data from each process on the line, but their MES system doesn’t have the functionality to deliver or the capacity to ingest it.

An MP&QM system is designed to take data collection and analysis to a new level, for full birth history trace the ability of every part that carries a serial number. This is accomplished by capturing the complete waveform of a process – not a few, but thousands of data points per cycle.

Where SPC plays its role to provide a top-level overview of the performance of processes on the line, MP&QM captures the full “DNA” of each process.

From leak test, to rundown operations or sealant and adhesive dispensing,this provides immediate visibility into the impact of part and process variations, and the interaction between processes. It allows for parts to be re-tested in a virtual digital environment using the source sensor data, selectively recall parts by applying a library of mathematical functions to the raw waveform, and so on.

But on a manufacturing line that may be rolling out thousands, even tens of thousands of units a day, collecting and managing this much data creates a whole new set of challenges.

To go that layer deeper than a traditional MES system, manufacturers need to take better care of their data.

But is your data ready for MP&QM?

To adopt an MP&QM system, it’s not enough to have C-levelapprovalfor the infrastructure and IT purchases, or the support of the rank and file. If data is missing, inaccurate or not delivered in an easily usable form, confidence can be lost, the initiative can fail and the investment will be wasted.

The minute a data stream starts to flow, there should be an “aha” moment among users as they see data in “real time” (which we define as within one cycle of process completion).

Three critical factors must be addressed to accomplish this:

1. Ensure the quality and content of data.Make sure your system’s judgement criteria are clearly defined and instrumented in the data stream. You need to define and understand what you are testing and the type and volume of data this requires.Maybe not all your production line processes require waveform analysis, or the same degree of granularity.

2. Break down the silos. We’ve seen one major automaker take two weeks to track down the full birth history of a single vehicle because the data is scattered among 15 different databases. While it is possible to “kluge” data after the fact, it just makes more sense if the data is transformed at source into a good model that allows for easily correlation to understand causal effect.

3. Build a good model. Take the time to structure an intuitive data requirements model with clear specifications and standards that define what is a pass and what is a fail. This can help drive the process of implementing an MP&QM system in its infancy, even at the purchasing stage, so data collection can start even before the new equipment hits the plant floor. If it’s already on your plant floor, you have an accepted standard to go by.

Yes, the future is now

Manufacturers must embrace a culture of continuous improvement to meet the challenge of Industry 4.0. This requires investment in technologies that take data collection, analysis and visualization one layer deeper than conventional MES, to boost first-time yields, reduce waste and ensure more consistent product quality.

Supplementing your MES investment with a robust MP&QM system is well worth it. We’ve seen first-time yield percentages jump from the low 80s to the high 90s. The cost savings and benefits that come from this magnitude of improvement have a huge impact on the bottom line.

There has already been a tenfold increase in the past few years in the amount of data stored per part in industries such as automotive and off-highway. Manufacturers that attempt to ride this big data wave without the right tools risk being drowned by it.

Weekly Brief

10 Most Promising Industrial Adhesives Solutions Companies - 2020

Read Also

Digital Disruption at a large Property Development company

Digital Disruption at a large Property Development company

Robyne Evans, Group Head of Knowledge Management, Goodman [ASX: GMG] Chris Mullan, Regional Head of Knowledge Management, Goodman [ASX: GMG]
Cultivating Fit-for-Purpose AI

Cultivating Fit-for-Purpose AI

Gautam Aggarwal, Senior Vice President, Operations & Technology, Asia Pacific, Mastercard (ASX – DHG)
Winning in the race of efficiency: Semi Plug & Play AI Services

Winning in the race of efficiency: Semi Plug & Play AI Services

Pooyan Asgari, Chief Data Officer, Domain Group [ASX – DHG]
With Digital Transformation, it might be time to manage

With Digital Transformation, it might be time to manage "what success looks like" differently

Penny Murphy, Regional Head of Digital Transformation, Asia,Arcadis [Euronext: ARCAD]
Robots, 3D and AI: Tech Trends Impacting the Apparel Supply Chain

Robots, 3D and AI: Tech Trends Impacting the Apparel Supply Chain

Marcus Chung, VP, Manufacturing and Supply Chain, ThirdLove
Submarine-propellers: big challenge for the

Submarine-propellers: big challenge for the "machining twins"

Elena Schmidt-Schmiedebach, Marketing Lead, North America, Starrag USA Inc.