How machine learning can improve Individual Section Machine effectiveness

individual-section-machine-process

It takes just 11 seconds to turn molten glass into a bottle. The most advanced hollow glass-forming operations can melt up to 300 tons of glass a day and produce around 50,000 bottles per hour.

Glass container manufacturing is only economical – and profitable – if it’s done on a large scale. Interrupting this high-speed process for maintenance or troubleshooting can be expensive, and process deviation can generate much scrap before the problem is identified and corrective action is taken.

The glass-forming process is precision manufacturing at its finest. Furnaces burn continuously at temperatures as high as 2800ºF. Precisely timed blades cut molten glass into gobs of an exact weight and size. And the viscosity has to be just right before the glass is blown into shape.

It’s more economical to run furnaces and Individual Section Machines (ISMs) 24/7, because thermal cycling ages equipment and takes time and energy to return to optimal temperatures. This means there’s little room for downtime, and planned stops must be tightly managed. Improving overall equipment effectiveness is crucial if glass manufacturers are to keep up with rising global demand, with the glass packaging market expected to register a compound annual growth rate (CAGR) of 4.39%, to reach $73 billion by 2026, up from $56 billion in 2020.

The increase is being driven by the conscious consumer. As one of the most trusted forms of packaging, glass is recyclable, better for our health than plastic, and preserves the flavor and freshness of food and beverages.

So, to truly compete in the Fourth Industrial Revolution, glass manufacturers must prioritize performance and flexibility. Ironically, the Fourth Industrial Revolution is a key driver to help glassmakers optimize glass forming processes and drive profitable growth.

End-to-end efficiency

The glass-forming process comprises several steps: melting, shaping, annealing, and cooling. There are many interactions between these processes and more than 150 variables that require a high level of expertise to troubleshoot when something goes wrong.

Individual Section Machines require constant operator care. Generally, work within an ISM falls into three categories: job change-over (to set up the machines for a particular run); routine operation; and maintenance.

Depending on the desired properties of the end product, machine operators must adjust the raw material recipe and machine setpoints between batches. The constant changes make it difficult to maintain an optimal glass-forming process from hot end forming, to cold end inspection. That means there’s a lot to get right and a lot more that could go wrong: the rate of having to make compensatory changes far exceeds the rate of human communication. Even real-time fixes are too late because it is fundamentally reactive.

Advanced artificial intelligence-powered systems and algorithms help factory managers to identify issues ahead of real-time. They accurately predict what will occur down the production line if specific actions are not taken immediately. Then, it goes a step further, to prescribe the machine setpoint adjustments that will avoid the undesirable outcome.

The glass production industry is well-positioned for artificial intelligence (AI) integration. As a mostly automated process, there’s an opportunity to optimize glass-forming even further with Expert Execution Systems (EES).

Optimizing Individual Section Machine effectiveness with EES

Individual Section Machines already have the capability to troubleshoot, repair, or maintain one section of the machine while the other sections keep working. To do this, the implicated section must first cool down. When the molten glass is reintroduced to the section, glass containers will be misshapen – and scrapped – until the optimal temperature is restored. Cooled glass can also get stuck in the machines, resulting in downtime and safety hazards as the glass has to be manually dislodged.

An Expert Execution System like DataProphet PRESCRIBE offers a faster, more efficient way to perform root cause analysis to improve production line performance, while reducing the time required for job changes and quality inspection. It monitors and optimizes key glass-forming parameters, such as machine speed, gob shape, and temperature, plunger contact time, cooling level, blow pressure, and blow time.

DataProphet PRESCRIBE ingests and analyzes historical process data and related quality data from every furnace, machine, and process. At the same time, it continually monitors, analyzes, and optimizes parameters and machine setpoints.

Using continuous AI-powered feedback loops between quality control and production teams, the system discovers the optimal recipe across the different subsets. It then prescribes the ideal machine setpoint ranges for the key controllable parameters. 

The EES’s algorithms are so advanced that it can alert managers via SMS before an anomaly is produced—PRESCRIBING corrective actions in anticipation of poor production states. It also tells the production team exactly what to do, to maintain production efficiency and keep machines within prescribed ranges.

What’s more, it suggests the best setpoints for each type of glass container across furnaces, dies and ISMs and prescribes changes across the production line to maintain optimum performance levels.

Ultimately, Expert Execution Systems improve Overall Equipment Effectiveness (OEE) by:

  • Improving availability by reducing unplanned stops due to quality troubleshooting, equipment failure, maintenance, and change over.
  • Improving performance by optimizing the production rate and making regular, small setpoint adjustments.
  • Improving quality through reduced scrap rate, rework, and returns.

With a stable environment, glass manufacturers can maintain low scrap levels and reduce the risk of interruptions to ISM sections. Plus, there are collateral benefits, such as lower cost of ownership, higher productivity, less reliance on operator expertise, fewer defects, and enhanced predictability of supply chain delivery.

When it comes to quality, an Expert Execution System provides high levels of traceability and supports faster inspection against 100 possible defect types, including spikes, blisters, cracks, and scratches.

The glass factory of the future will run on machine setpoints governed by AI. It will use integrated process control and advanced algorithms to modernize glass-forming factories and support them into the Fourth Industrial Revolution and beyond.


Find out more about how we can improve OEE for your IS Machine.

Leave a Reply

Your email address will not be published. Required fields are marked *