How to bring sensing in manufacturing out of the lab and onto the line

by Jamie Jeffs, Head of Industrial Instrumentation at 42T

A sensing revolution is under way in manufacturing. The advanced capability, accessibility and reduced cost of sensing technologies, such as vision systems and optical sensing techniques, have the potential to transform process control in the manufacture of high-volume products

For manufacturers, the benefits are obvious.

The measurable efficiency gains from enhanced quality control, improved manufacturing yield and reduced wastage are being demanded by industrial manufacturers keen to bring materials analysis and quality assurance out of the lab and onto the production line. Enabling high-precision measurement systems to be installed directly on to production lines providing continuous and real-time process data at key stages of the production process is an attractive proposition.  

Whilst for sensor and vision system providers this poses a potentially exciting opportunity for growth, it also poses a potential threat for those companies not able to offer the capability of these systems to meet this growing demand.  

By leveraging advances in new technology and addressing early in any development programme the key design challenges, many more sensor and instrumentation companies can join the sensing revolution and achieve the kind of transformational affect envisaged by large scale manufacturers.

The food and beverage industry is a good example of where there is both the drive for manufacturing innovation as well as challenges in delivering online sensing capabilities.  

In large-scale food and beverage manufacturing, quality control is crucial for maintaining high yield and throughput. Traditionally, human inspection and batch testing are used, but they may not catch all defects. Defective product that escapes the factory can have major impacts for the manufacturer, including the risk of product recalls, costly fines from regulators, and damage to reputation. 

Lower-cost vision systems are widely used for visual defect detection, however, achieving more detailed analysis of products on the line can be achieved thanks to advances in spectral analysis, such as Ultraviolet (UV), Short Wave Infrared (SWIR) and hyperspectral imaging. Developments in the capabilities of typically lower cost solutions such as imaging radar, ultrasound waveform analysis and time of flight imaging are also being used in a range of manufacturing processes. Leveraging these techniques is enabling material analysis to detect impurities in both raw materials and end  product online in production processes. Additionally, developments in both the responsiveness and accuracy of photoelectric and 3D scanning sensors have enabled automated finished goods inspection in real time.  

However, the practicalities of installing high-precision instruments in these processes can be complex. With such a vast range of varying applications, even within the same production process, the cost of implementation, integration and configuration of more complex measurement techniques remains a barrier to large-scale adoption of these technologies. 

Whilst sophisticated metrology systems can accurately inspect and verify conformance of a vast range of materials, the comparatively uncontrolled environment of the production line (compared to a controlled laboratory) poses challenges to high-precision measurement systems both in terms of maintaining sensor performance, and robustness. Aside from the cost of installing multiple sensors through a production process, the solutions are required to have a low level of maintenance and must work reliably. If processes are dependent on sensors performing production line quality assurance, a faulty sensor could equal costly plant downtime until it is fixed or replaced. 

This is even more important when sensing systems are installed to support automated process such a robotic handling, often used at the end of a production line in packaging and transporting finished goods from the factory. Interoperability of sensing solutions is also required to enable this integrated approach. More and more manufacturers are demanding that these capabilities are embedded within these sensor systems rather than be burdened with large-scale integration projects at significant additional cost.  

For companies wanting to take advantage of this sensing revolution, here are five key design and development considerations that are crucial to address early. 

1. Technology selection - understanding the application 

Firstly, there needs to be an understanding of the application that these solutions will serve and how developing an online solution will benefit the manufacturer. Large-scale adoption is unlikely if the solution being offered does not add significant value over existing methods e.g. lab testing or manual inspection. 

In transferring sensing and analysis out of the lab and onto the line, it should also not be assumed that taking an existing sensor technology and transferring this to a manufacturing process provides the best solution for the application.  

The same level of precision currently performed through batch-testing may not be necessary with continuous monitoring of 100% of product. In many cases, a simple pass/fail criteria may be the most valuable output. 

One example is radar imaging. By utilising a robust and typically lower cost technology (when compared to x-ray), radar imaging devices can be mounted directly onto the production line and provide sufficient imaging inside packaged goods with concealed materials (e.g. sealed food and beverage products or blister packs) and immediately alert operators to defects such as missing items, unfilled containers or damaged concealed primary packaging. 

By considering alternative technologies versus traditional methods, companies can potentially unlock opportunities in a wider range of applications. 

A robust assessment of the feasibility of deploying alternative measurement and detection systems will always be needed with a focus on managing the technical risks inherent in new technology development. However, with the correct approach, new technology deployment can enable companies to capitalise on the sensing revolution with a disruptive and scalable offering.  

2. Human factors - ease of use

When considering the development of solutions designed to be used in fast-moving production lines, the human factors of the people who will be using these systems, the insights they provide and the way they will interact with them are an important part of the design process. 

When high-precision measurement and detection is required, such as in the precision engineering industry, there aren’t just the technical challenges to consider, there is also the requirement to make these systems easy to integrate, easy to configure and capable of reliably determining the pass/fail criteria of the material it is inspecting. 

One example where high-precision defect detection coupled with ease of integration and ease of use is becoming more achievable for deployment in manufacturing environments is through the use of Machine Vision.  

This can be achieved using lightweight machine learning capabilities to create simple algorithms that allows imaging of a wide range of materials and objects directly on the production line.  

By training systems to determine what-good-looks-like and alert users to any object that deviates from this requires only a few examples of each object to be used to train the machine vision system. This avoids complicated configuration and set up by the users whilst still providing the pass/fail criteria built into the design of the system.

3. Environmental factors - low power, low maintenance and reliable performance

Material inspection tools moving out of a tightly controlled environment and into production lines with fast-moving product requires system engineered solutions that can account for multiple environmental factors. These factors vary depending on the technology used within the system and the applications they are intended to be used with. 

With all high-precision optical measurement techniques, managing impacts of variable temperature, humidity, and other environmental factors, such as ambient light & vibration is crucial. It’s a challenge system providers have to overcome in the design and development of these solutions. 

Equally the restrictions on energy usage and power consumption demands that sensor providers consider low power solutions that do not require access to mains power in plant. 

Conditions within production processes also increase risks of contamination of sensitive optical surfaces that can affect performance and require more frequent maintenance. 

Advances in microelectronics are enabling sensor manufacturers to add multiple sensor inputs to their measurement devices to dynamically address impacts on precision instruments without increasing the size and weight of online measurement systems. An example includes the use of micro resistance temperature devices (RTDs). Continuous monitoring of temperature within optical sensors enables changes in laser characteristics to be dynamically adjusted based on the known impacts of fluctuating temperature. 

Advances in ultra-lower power electronics also provides opportunities to drive precision instruments via low power means. 

Interesting developments in surface coatings providing self-cleaning effects on glass or optical surfaces have the potential to provide an effective solution to rapid contamination. Reducing the rate of contamination when deployed in a manufacturing environment would enable less frequent maintenance and improve the robustness of precision sensors.  

Ensuring system design incorporates requirements for low-power, low-maintenance and robust, reliable performance is key. Assessing and designing in these capabilities gives the best opportunity to manage the wide range of environmental factors that sensing on the line has to address. 

4. Design modularity - a common hardware platform

The ability to provide sensing solutions across such a wide range of industries requires a solution that can adapt to a wide variety of applications. In many production processes, this includes different parameters needed within the sensors for different products being produced on the same production line. A consequence of this approach is that solution providers often end up having to manage many discrete variants of a single product or fail to offer solutions to the widest range of applications. 

Designing solutions based on a common hardware platform can address these issues but this often requires a reliance on programmable firmware and software to cope with the wide variety of possible applications. For high precision sensing in manufacturing process, this can be limited by the capacity to store and programme multiple parameters on a single device. 

With the continuous advancement in microprocessor speed and capacity sensing companies can now seek to achieve a greater level of modular design utilising greater processing embedded within their devices. 

An example of this can be seen with 3D imaging sensors via laser triangulation. This technique requires the user to define multiple areas of interest with each object. Previously, this has been limited by the system capacity to inspect at the required level of tolerance and speed of operation. With faster processing speeds and greater capacity for storing multiple configurations, this can now be achieved for online measurement across multiple products using a common hardware platform. 

5. Interoperability - automation and connectivity

Manufacturers increasingly expect sensor companies to provide a fully integrated solution that can be easily deployed. To achieve this, it is important that the design intent for automation of these sensors is clearly defined and designed into any sensor development, along with connectivity, and interoperability with other plant systems. 

This includes the ability to work collaboratively with robotic systems ('cobots') for correct placement and positioning to achieve the precision inspection required, interoperability within PLC systems to trigger action when a defect is detected or developing the algorithms and machine learning tools required to rapidly learn and identify defective product.   

This connectivity is required for manufacturers to achieve the improved productivity and quality assurance, online sensing can provide. 

Historically, the capability to deliver these solutions has sat with a few major automation specialists and integrators providing the gateway between sensing and large plant. However, despite the wide array of industrial input and output options used across manufacturing, solutions now exist that enable instrumentation companies to embed these capabilities within their systems. 

It is generally accepted that connectivity in process control sensors must remain open for users to connect to other devices and PLC systems on site. However, increasingly open-source protocols, such as IO-Link are being adopted by sensor companies along with required gateway devices and wireless communication systems expected from manufacturers. 

System providers who are able to offer this integration, in a user-friendly way, having designed this into their final released offering, will have a distinct advantage over the competition. Even when using opensource solutions, the implementation and support tools required to enable simple integration and interoperability with PLCs, automated processes and data management software have to be carefully considered in the design process. 

Towards larger scale adoption

The potential benefits of sensing solutions in high-volume manufacturing processes have been understood for many years. Now with advances in core measurement technologies and materials science, high reliability electronics, interconnectivity, automation and user-centric design, the opportunity exists for large scale adoption in manufacturing. 

For companies wanting to be at the forefront of this sensing revolution, providing fully integrated systems to service the growing demand is a major strategic decision. It is therefore critical to tackle the challenging requirements of online continuous measurement and detection for these environments early in the design and development process. 

Companies able to offer reliable, affordable, low-maintenance and easy-to-use sensing have opportunities to access new markets and generate new revenue streams. Longer term, this enables measurement specialists to cement their market position as industry leaders offering manufacturers critical insights to analyse and improve product quality and operational efficiency. 

At 42T we work closely with both high-volume manufacturers looking to implement these solutions and innovative instrumentation companies in the development of these fully integrated sensing systems. We are therefore well placed to support product development of these solutions at all stages of the development process with the capabilities to address the complex technical challenges of taking sensing out of the lab and onto the line. 

Read about our work here.

About Jamie

Jamie is an experienced industrial products specialist having held senior roles within market-leading instrumentation businesses. He has led teams in the design, development, manufacturing, sales and marketing of industrial measurement systems for environmental and process control.

Working with some of the world's largest companies in power generation and manufacturing process, he has provided instrumentation solutions for clients across a range of technologies often for use in the harsh environment of heavy industry.

If you would like to find out more, please contact Jamie Jeffs:
answers@42T.com | +44 (0)1480 302700 | Jamie Jeffs | LinkedIn