Walk onto many plastics processors’ shop floors today, and you’re likely to see a mix of smart machines and legacy equipment. Ideally those older systems have been upgraded with manufacturing network communication boards or external sensors connected to a monitoring device to obtain real-time insights from all their machinery.

However, some plastics manufacturers never got past the hurdles of using older versions of these machines, sensors and traditional connectivity options, leaving them with knowledge gaps that can lead to unnecessary scrap, missed deadlines or worse.

The good news is that a new generation of lower-cost smart sensors and industry-standard connectivity protocols are making it easier, faster and more affordable to augment legacy shop floor equipment. As a result, plastics manufacturers can more readily track the performance and quality metrics that help address potential issues early, control costs, optimize resource utilization and automate manufacturing workflows.

In Part 1 of this two-article series, we will begin with a review of relevant sensors for plastics manufacturing. Then we’ll examine how new smart sensors that support industry communications protocols are making it significantly easier to extract meaningful data from legacy machines to drive efficiency, quality, and cost savings.

Common sensors in plastics processing

Available sensors on the market support different types of data collection and enable plastics processors to track various metrics related to productivity, consumption, wear and other factors to monitor and improve their production operations. Here are some of the most widely used.

Proximity sensors can be placed on machines to capture movement, which can then be associated with multiple stages of manufacturing. One sensor may capture parts produced. Another may be added to a hopper’s motor wheel to associate wheel rotation with pellet volume. Comparing those two measurements can, for example, help the team understand their scrap and percentage of loss.

Amperage sensors are used to continuously monitor the electrical load of machines like extruders, injection molding presses and granulators, which provides insights into machine speed and force of tooling. This helps process engineers relate machine amperage to screw movement and mold closing, which can be turned into process times and potentially pressures to compare with part recipes — enabling real-time quality part assessment.

Pressure sensors can take multiple forms in plastics processing. For example, sensors inside a mold cavity measure the pressure of the plastic as it fills and cools. Other sensors track hydraulic pressure powering a machine or injection cylinder to ensure the exact amount of force is used. Melt pressure sensors ensure that the molten plastic is forced through the die consistently. Similarly, in blow molding, pressure sensors regulate the exact air pressure used to inflate the plastic parison. Taken together, they can help identify issues before they significantly affect part quality and equipment maintenance.

Vibration sensors measure the amount and frequency of vibration in a given machine or piece of equipment. Those measurements, in turn, can be used to detect imbalances and other issues to predict maintenance issues and avoid unplanned downtime. Additionally, having a vibration base measurement while making parts can provide a machine signature for good parts and a healthy machine.

Flow meters can be added to devices to track fluid-based processes in additive injection lines, polymer melt delivery and mold cooling systems. Using flow meters helps to ensure optimal cooling and consistent product quality while minimizing waste.

Humidity sensors are used on equipment, such as hoppers and extruders, for real-time monitoring of a resin’s moisture content. This helps to prevent defects like weak structural integrity, bubbling and brittleness.

Historically, these sensors were fairly simple. They would produce counts or measures at pre-set intervals, and they typically lacked built-in communications support. So, an intermediary programmable logic controller (PLC) was needed to capture data from sensors; convert it to a usable computer format; and make it available to manufacturing execution system (MES), enterprise resource planning (ERP) and other software used to run manufacturing operations.

In short, extracting value from traditional sensors could be complex, time-consuming and expensive. But that is changing with newer alternatives.

Next-generation sensors simplify access to insights

In the last few years, we have seen a growing number of companies introduce next-generation sensors that feature built-in artificial intelligence (AI) and support for widely adopted networking communications protocols. These, coupled with improved wireless connectivity, are lowering the cost and complexity of deriving meaningful insights from legacy machine data.

The embedded AI in sensors is generally basic machine learning that resides on a chip, but it allows plastics manufacturers to gain relevant insights right from the sensor itself. Take the example of vibration sensors. There’s not much value from just seeing a six- or seven-kilohertz readout of the machine vibration at a given time. However, newer, AI-enabled vibration sensors can be set up to capture when a vibration is more intense than usual, potentially signaling an issue that impacts machine wear or the parts being produced.

Meanwhile, built-in support for communications protocols has removed the need for intermediary PLCs to extract data from sensors into a plastics manufacturer’s software. It has also made it easier to mix-and-match different sensors on the shop floor. The two most widely adopted protocols are Message Queuing Telemetry Transport (MQTT) and Open Platform Communications Unified Architecture (OPC UA).

MQTT is an Oasis standard messaging protocol designed for machine-to-machine (M2M) communications and the Internet of Things (IoT). Two qualities of MQTT make it ideal for supporting manufacturing operations. First, its light footprint works well for small devices like sensors, even those running within unreliable networks. Second, its publish/subscribe model removes the need to know the internet protocol (IP) address of another device or machine, simplifying set-up. Newer sensors that support MQTT serve as the publisher.