In the pursuit of sustainability, sensors are reducing cycle times, energy use and waste, automating closed-loop process control and increasing knowledge, opening up new possibilities for smart manufacturing and structures.#sensors #sustainability #SHM
Sensors on the left (top to bottom): heat flux (TFX), in-mold dielectrics (Lambient), ultrasonics (University of Augsburg), disposable dielectrics (Synthesites) and between pennies and thermocouples Microwire (AvPro).Graphs (top, clockwise): Collo dielectric constant (CP) versus Collo ionic viscosity (CIV), resin resistance versus time (Synthesites) and digital model of caprolactam implanted preforms using electromagnetic sensors (CosiMo project, DLR ZLP , University of Augsburg).
As the global industry continues to emerge from the COVID-19 pandemic, it has shifted to prioritizing sustainability, which requires reducing waste and consumption of resources (such as energy, water and materials).As a result, manufacturing must become more efficient and smarter.But this requires information.For composites, where does this data come from?
As described in CW’s 2020 Composites 4.0 series of articles, defining the measurements needed to improve part quality and production, and the sensors needed to achieve those measurements, is the first step in smart manufacturing.During 2020 and 2021, CW reported on sensors—dielectric sensors, heat flux sensors, fiber optic sensors, and non-contact sensors using ultrasonic and electromagnetic waves—as well as projects demonstrating their capabilities (see CW’s online sensor content set).This article builds on this report by discussing the sensors used in composite materials, their promised benefits and challenges, and the technological landscape under development.Notably, companies that are emerging as leaders in the composites industry are already exploring and navigating this space.
Sensor network in CosiMo A network of 74 sensors – 57 of which are ultrasonic sensors developed at the University of Augsburg (shown on the right, light blue dots in the upper and lower mould halves) – are used for Lid demonstrator for the T-RTM molding CosiMo project for thermoplastic composite batteries.Image credit: CosiMo project, DLR ZLP Augsburg, University of Augsburg
Goal #1: Save money.The CW’s December 2021 blog, “Custom Ultrasonic Sensors for Composite Process Optimization and Control,” describes work at the University of Augsburg (UNA, Augsburg, Germany) to develop a network of 74 sensors that For the CosiMo project to manufacture an EV battery cover demonstrator (composite materials in smart transportation).The part is fabricated using thermoplastic resin transfer molding (T-RTM), which polymerizes caprolactam monomer in situ into a polyamide 6 (PA6) composite.Markus Sause, Professor at UNA and Head of UNA’s Artificial Intelligence (AI) Production Network in Augsburg, explains why sensors are so important: “The biggest advantage we offer is the visualization of what is happening inside the black box during processing. Currently, most manufacturers have limited systems to achieve this. For example, they use very simple or specific sensors when using resin infusion to make large aerospace parts. If the infusion process goes wrong, you basically have a big piece of scrap. But if you have a solution solutions to understand what went wrong in the production process and why, you can fix it and correct it, saving you a lot of money.”
Thermocouples are an example of a “simple or specific sensor” that has been used for decades to monitor the temperature of composite laminates during autoclave or oven curing.They are even used to control the temperature in ovens or heating blankets to cure composite repair patches using thermal bonders.Resin manufacturers use a variety of sensors in the lab to monitor changes in resin viscosity over time and temperature to develop cure formulations.What is emerging, however, is a sensor network that can visualize and control the manufacturing process in situ based on multiple parameters (eg, temperature and pressure) and the state of the material (eg, viscosity, aggregation, crystallization).
For example, the ultrasonic sensor developed for the CosiMo project uses the same principles as ultrasonic inspection, which has become the mainstay of non-destructive testing (NDI) of finished composite parts.Petros Karapapas, Principal Engineer at Meggitt (Loughborough, UK), said: “Our aim is to minimise the time and labour required for post-production inspection of future components as we move towards digital manufacturing.” Materials Centre (NCC, Bristol, UK) collaboration to demonstrate the monitoring of a Solvay (Alpharetta, GA, USA) EP 2400 ring during RTM using a linear dielectric sensor developed at Cranfield University (Cranfield, UK) Flow and curing of oxyresin for a 1.3 m long, 0.8 m wide and 0.4 m deep composite shell for a commercial aircraft engine heat exchanger.“As we looked at how to make larger assemblies with higher productivity, we couldn’t afford to do all the traditional post-processing inspections and testing on every part,” Karapapas said.”Right now, we make test panels next to these RTM parts and then do mechanical testing to validate the cure cycle. But with this sensor, that’s not necessary.”
The Collo Probe is immersed in the paint mixing vessel (green circle at the top) to detect when mixing is complete, saving time and energy.Image credit: ColloidTek Oy
“Our goal is not to be another laboratory device, but to focus on production systems,” says Matti Järveläinen, CEO and founder of ColloidTek Oy (Kolo, Tampere, Finland).The CW January 2022 blog “Fingerprint Liquids for Composites” explores Collo’s combination of electromagnetic field (EMF) sensors, signal processing and data analysis to measure the “fingerprint” of any liquid such as monomers, resins or adhesives .“What we offer is a new technology that provides direct feedback in real time, so you can better understand how your process is actually working and react when things go wrong,” says Järveläinen.“Our sensors convert real-time data into understandable and actionable physical quantities, such as rheological viscosity, which allow process optimization. For example, you can shorten mixing times because you can clearly see when mixing is complete. Therefore, with You can increase productivity, save energy and reduce scrap compared to less optimized processing.”
Goal #2: Increase process knowledge and visualization.For processes like aggregation, Järveläinen says, “You don’t see much information from just a snapshot. You’re just taking a sample and going into the lab and looking at what it was like minutes or hours ago. It’s like driving on the highway, every hour Open your eyes for a minute and try to predict where the road is going.” Sause agrees, noting that the sensor network developed in CosiMo “helps us get a complete picture of the process and material behavior. We can see local effects in the process, in response to Variations in part thickness or integrated materials such as foam core. What we are trying to do is provide information about what is actually happening in the mold. This allows us to determine various information such as the shape of the flow front, the arrival of each part time and the degree of aggregation at each sensor location.”
Collo works with manufacturers of epoxy adhesives, paints and even beer to create process profiles for each batch produced.Now every manufacturer can view the dynamics of their process and set more optimized parameters, with alerts to intervene when batches are out of specification.This helps stabilize and improve quality.
Video of the flow front in a CosiMo part (injection entrance is the white dot in the center) as a function of time, based on measurement data from an in-mold sensor network.Image credit: CosiMo project, DLR ZLP Augsburg, University of Augsburg
“I want to know what happens during part manufacture, not open the box and see what happens afterward,” says Meggitt’s Karapapas.”The products we developed using Cranfield’s dielectric sensors allowed us to see the in-situ process, and we were also able to verify the curing of the resin.” Using all six types of sensors described below (not an exhaustive list, just a small selection, suppliers, too), can monitor cure/polymerization and resin flow.Some sensors have additional capabilities, and combined sensor types can expand the tracking and visualization possibilities during composite molding.This was demonstrated during CosiMo, which used ultrasonic, dielectric and piezoresistive in-mode sensors for temperature and pressure measurements by Kistler (Winterthur, Switzerland).
Goal #3: Reduce cycle time.Collo sensors can measure the uniformity of two-part fast-curing epoxy as parts A and B are mixed and injected during RTM and at every location in the mold where such sensors are placed.This could help enable faster cure resins for applications such as Urban Air Mobility (UAM), which would provide faster cure cycles compared to current one-part epoxies such as RTM6.
Collo sensors can also monitor and visualize epoxy being degassed, injected and cured, and when each process is complete.Finishing curing and other processes based on the actual state of the material being processed (versus traditional time and temperature recipes) is called material state management (MSM).Companies such as AvPro (Norman, Oklahoma, USA) have been pursuing MSM for decades to track changes in part materials and processes as it pursues specific targets for glass transition temperature (Tg), viscosity, polymerization and/or crystallization .For example, a network of sensors and digital analysis in CosiMo were used to determine the minimum time required to heat up the RTM press and mold and found that 96% of the maximum polymerization was achieved in 4.5 minutes.
Dielectric sensor suppliers such as Lambient Technologies (Cambridge, MA, USA), Netzsch (Selb, Germany) and Synthesites (Uccle, Belgium) have also demonstrated their ability to reduce cycle times.Synthesites’ R&D project with composites manufacturers Hutchinson (Paris, France) and Bombardier Belfast (now Spirit AeroSystems (Belfast, Ireland)) reports that based on real-time measurements of resin resistance and temperature, through its Optimold data acquisition unit and Optiview Software converts to estimated viscosity and Tg.“Manufacturers can see the Tg in real time, so they can decide when to stop the curing cycle,” explains Nikos Pantelelis, Director of Synthesites.“They don’t have to wait to complete a carryover cycle that is longer than necessary. For example, the traditional cycle for RTM6 is a 2-hour full cure at 180°C. We’ve seen that this can be shortened to 70 minutes in some geometries. This was also demonstrated in the INNOTOOL 4.0 project (see “Accelerating RTM with Heat Flux Sensors”), where the use of a heat flux sensor shortened the RTM6 cure cycle from 120 minutes to 90 minutes.
Goal #4: Closed-loop control of adaptive processes.For the CosiMo project, the ultimate goal is to automate closed-loop control during the production of composite parts.This is also the goal of the ZAero and iComposite 4.0 projects reported by CW in 2020 (30-50% cost reduction).Note that these involve different processes – automated placement of prepreg tape (ZAero) and fiber spray preforming compared to high pressure T-RTM in CosiMo for RTM with fast curing epoxy (iComposite 4.0).All of these projects use sensors with digital models and algorithms to simulate the process and predict the outcome of the finished part.
Process control can be thought of as a series of steps, Sause explained.The first step is to integrate sensors and process equipment, he said, “to visualize what’s going on in the black box and the parameters to use. The other few steps, maybe half of closed-loop control, are being able to push the stop button to intervene, Tune the process and prevent rejected parts. As a final step, you can develop a digital twin, which can be automated, but also requires investment in machine learning methods.” In CosiMo, this investment enables sensors to feed data into the digital twin, Edge analysis (calculations performed at the edge of the production line versus calculations from a central data repository) is then used to predict flow front dynamics, fiber volume content per textile preform and potential dry spots.”Ideally, you can establish settings to enable closed-loop control and tuning in the process,” Sause said.”These will include parameters like injection pressure, mold pressure and temperature. You can also use this information to optimize your material.”
In doing so, companies are using sensors to automate processes.For example, Synthesites is working with its customers to integrate sensors with equipment to close the resin inlet when infusion is complete, or turn on the heat press when target cure is achieved.
Järveläinen notes that to determine which sensor is best for each use case, “you need to understand what changes in the material and process you want to monitor, and then you have to have an analyzer.” An analyzer acquires the data collected by an interrogator or data acquisition unit. raw data and convert it into information usable by the manufacturer.”You actually see a lot of companies integrating sensors, but then they don’t do anything with the data,” Sause said.What is needed, he explained, is “a system of data acquisition, as well as a data storage architecture to be able to process the data.”
“End users don’t just want to see raw data,” says Järveläinen.”They want to know, ‘Is the process optimized?’” When can the next step be taken?”To do this, you need to combine multiple sensors for analysis, and then use machine learning to speed up the process.” This edge analysis and machine learning approach used by the Collo and CosiMo team can be achieved through viscosity maps, numerical models of the resin flow front, and The ability to ultimately control process parameters and machinery is visualized.
Optimold is an analyzer developed by Synthesites for its dielectric sensors.Controlled by Synthesites’ Optiview software, the Optimold unit uses temperature and resin resistance measurements to calculate and display real-time graphs to monitor resin status including mix ratio, chemical aging, viscosity, Tg and degree of cure.It can be used in prepreg and liquid forming processes.A separate unit Optiflow is used for flow monitoring.Synthesites has also developed a curing simulator that does not require a curing sensor in the mold or part, but instead uses a temperature sensor and resin/prepreg samples in this analyzer unit.“We are using this state-of-the-art method for infusion and adhesive curing for wind turbine blade production,” said Nikos Pantelelis, Director of Synthesites.
Synthesites process control systems integrate sensors, Optiflow and/or Optimold data acquisition units, and OptiView and/or Online Resin Status (ORS) software.Image credit: Synthesites, edited by The CW
Therefore, most sensor suppliers have developed their own analyzers, some using machine learning and some not.But composite manufacturers can also develop their own custom systems or buy off-the-shelf instruments and modify them to meet specific needs.However, analyzer capability is only one factor to consider.There are many others.
Contact is also an important consideration when choosing which sensor to use.The sensor may need to be in contact with the material, the interrogator, or both.For example, heat flux and ultrasonic sensors can be inserted into an RTM mold 1-20mm from the surface – accurate monitoring does not require contact with the material in the mold.Ultrasonic sensors can also interrogate parts at different depths depending on the frequency used.Collo electromagnetic sensors can also read the depth of liquids or parts – 2-10 cm, depending on the frequency of interrogation – and through non-metallic containers or tools in contact with the resin.
However, magnetic microwires (see “Non-contact monitoring of temperature and pressure inside composites”) are currently the only sensors capable of interrogating composites at a distance of 10 cm.That’s because it uses electromagnetic induction to elicit a response from the sensor, which is embedded in the composite material.AvPro’s ThermoPulse microwire sensor, embedded in the adhesive bond layer, has been interrogated through a 25mm thick carbon fiber laminate to measure temperature during the bonding process.Since the microwires have a hairy diameter of 3-70 microns, they do not affect composite or bondline performance.At slightly larger diameters of 100-200 microns, fiber optic sensors can also be embedded without degrading structural properties.However, because they use light to measure, fiber optic sensors must have a wired connection to the interrogator.Likewise, since dielectric sensors use voltage to measure resin properties, they must also be connected to an interrogator, and most must also be in contact with the resin they are monitoring.
The Collo Probe (top) sensor can be immersed in liquids, while the Collo Plate (bottom) is installed in the wall of a vessel/mixing vessel or process piping/feed line.Image credit: ColloidTek Oy
The temperature capability of the sensor is another key consideration.For example, most off-the-shelf ultrasonic sensors typically operate at temperatures up to 150°C, but parts in CosiMo need to be formed at temperatures above 200°C.Therefore, UNA had to design an ultrasonic sensor with this capability.Lambient’s disposable dielectric sensors can be used on part surfaces up to 350°C, and its reusable in-mold sensors can be used up to 250°C.RVmagnetics (Kosice, Slovakia) has developed its microwire sensor for composite materials that can withstand curing at 500°C.While the Collo sensor technology itself has no theoretical temperature limit, the tempered glass shield for the Collo Plate and the new polyetheretherketone (PEEK) housing for the Collo Probe are both tested for continuous duty at 150°C, according to Järveläinen.Meanwhile, PhotonFirst (Alkmaar, The Netherlands) used a polyimide coating to provide an operating temperature of 350°C for its fiber optic sensor for the SuCoHS project, for a sustainable and cost-effective high-temperature composite.
Another factor to consider, especially for installation, is whether the sensor measures at a single point or is a linear sensor with multiple sensing points.For example, Com&Sens (Eke, Belgium) fiber optic sensors can be up to 100 meters long and feature up to 40 fiber Bragg grating (FBG) sensing points with a minimum spacing of 1 cm.These sensors have been used for structural health monitoring (SHM) of 66-meter-long composite bridges and resin flow monitoring during infusion of large bridge decks.Installing individual point sensors for such a project would require a large number of sensors and a lot of installation time.NCC and Cranfield University claim similar advantages for their linear dielectric sensors.Compared to single-point dielectric sensors offered by Lambient, Netzsch and Synthesites, “With our linear sensor, we can monitor resin flow continuously along the length, which significantly reduces the The number of sensors required in the part or tool.”
AFP NLR for Fiber Optic Sensors A special unit is integrated into the 8th channel of the Coriolis AFP head to place four fiber optic sensor arrays into a high temperature, carbon fiber reinforced composite test panel.Image credit: SuCoHS Project, NLR
Linear sensors also help automate installations.In the SuCoHS project, Royal NLR (Dutch Aerospace Centre, Marknesse) developed a special unit integrated into the 8th channel Automated Fiber Placement (AFP) head of Coriolis Composites (Queven, France) to embed Four arrays (separate fiber optic lines), each with 5 to 6 FBG sensors (PhotonFirst offers a total of 23 sensors), in carbon fiber test panels.RVmagnetics has placed its microwire sensors in pultruded GFRP rebar.”The wires are discontinuous [1-4 cm long for most composites microwires], but are automatically placed continuously when the rebar is produced,” said Ratislav Varga, co-founder of RVmagnetics. “You have a microwire with a 1km microwire. coils of filament and feed it into the rebar production facility without changing the way the rebar is made.” Meanwhile, Com&Sens is working on automated technology to embed fiber-optic sensors during the filament winding process in pressure vessels.
Because of its ability to conduct electricity, carbon fiber can cause problems with dielectric sensors.Dielectric sensors use two electrodes placed close to each other.”If the fibers bridge the electrodes, they short-circuit the sensor,” explains Lambient founder Huan Lee.In this case, use a filter.”The filter lets the resin pass the sensors, but insulates them from the carbon fiber.” The linear dielectric sensor developed by Cranfield University and NCC uses a different approach, including two twisted pairs of copper wires.When a voltage is applied, an electromagnetic field is created between the wires, which is used to measure resin impedance.The wires are coated with an insulating polymer that doesn’t affect the electric field, but prevents the carbon fiber from shorting out.
Of course, cost is also an issue.Com&Sens states that the average cost per FBG sensing point is 50-125 euros, which may drop to around 25-35 euros if used in batches (eg, for 100,000 pressure vessels).(This is only a fraction of the current and projected production capacity of composite pressure vessels, see CW’s 2021 article on hydrogen.) Meggitt’s Karapapas says he has received offers for fibre optic lines with FBG sensors averaging £250/sensor (≈300€/sensor), the interrogator is worth around £10,000 (€12,000).”The linear dielectric sensor we tested was more like a coated wire that you can buy off the shelf,” he added.”The interrogator we use,” adds Alex Skordos, reader (senior researcher) in Composites Process Science at Cranfield University, “is an impedance analyzer, which is very accurate and costs at least £30,000 [≈ €36,000], But the NCC uses a much simpler interrogator that basically consists of off-the-shelf modules from the commercial company Advise Deta [Bedford, UK].” Synthesites is quoting €1,190 for in-mold sensors and €20 for single-use/part sensors In EUR, Optiflow is quoted at EUR 3,900 and Optimold at EUR 7,200, with increasing discounts for multiple analyzer units.These prices include Optiview software and any necessary support, Pantelelis said, adding that wind blade manufacturers save 1.5 hours per cycle, add blades per line per month, and reduce energy use by 20 percent, with a return on investment of only for four months.
Companies using sensors will gain an advantage as composites 4.0 digital manufacturing evolves.For example, says Grégoire Beauduin, Director of Business Development at Com&Sens, “As pressure vessel manufacturers try to reduce weight, material usage and cost, they can use our sensors to justify their designs and monitor production as they reach the required levels by 2030. The same sensors used to assess strain levels within layers during filament winding and curing can also monitor tank integrity during thousands of refueling cycles, predict required maintenance and recertify at the end of design life. We can A digital twin data pool is provided for every composite pressure vessel produced, and the solution is also being developed for satellites.”
Enabling digital twins and threads Com&Sens is working with a composites manufacturer to use its fiber optic sensors to enable digital data flow through design, production and service (right) to support digital ID cards that support the digital twin of each part (left) made.Image credit: Com&Sens and Figure 1, “Engineering with Digital Threads” by V. Singh, K. Wilcox.
Thus, sensor data supports the digital twin, as well as the digital thread that spans design, production, service operations and obsolescence.When analyzed using artificial intelligence and machine learning, this data feeds back into design and processing, improving performance and sustainability.This has also changed the way supply chains work together.For example, adhesive manufacturer Kiilto (Tampere, Finland) uses Collo sensors to help its customers control the ratio of components A, B, etc. in their multi-component adhesive mixing equipment.”Kiilto can now adjust the composition of its adhesives for individual customers,” says Järveläinen, “but it also allows Kiilto to understand how resins interact in customers’ processes, and how customers interact with their products, which is changing how supply is made. Chains can work together.”
OPTO-Light uses Kistler, Netzsch and Synthesites sensors to monitor curing for thermoplastic overmolded epoxy CFRP parts.Image credit: AZL
Sensors also support innovative new material and process combinations.Described in CW’s 2019 article on the OPTO-Light project (see “Thermoplastic Overmolding Thermosets, 2-Minute Cycle, One Battery”), AZL Aachen (Aachen, Germany) uses a two-step process to horizontally compress a single To (UD) carbon fiber/epoxy prepreg, then overmolded with 30% short glass fiber reinforced PA6.The key is to only partially cure the prepreg so that the remaining reactivity in the epoxy can enable bonding to the thermoplastic.AZL uses Optimold and Netzsch DEA288 Epsilon analyzers with Synthesites and Netzsch dielectric sensors and Kistler in-mold sensors and DataFlow software to optimize injection molding.”You have to have a deep understanding of the prepreg compression molding process because you have to make sure you understand the state of cure in order to achieve a good connection to thermoplastic overmolding,” explains AZL research engineer Richard Schares. “In the future, the process may be adaptive And intelligent, process rotation is triggered by sensor signals.”
However, there is a fundamental problem, says Järveläinen, “and that is the lack of understanding by customers on how to integrate these different sensors into their processes. Most companies don’t have sensor experts.” Currently, the way forward requires sensor manufacturers and customers Exchange information back and forth.Organizations such as AZL, DLR (Augsburg, Germany) and NCC are developing multi-sensor expertise.Sause said there are groups within UNA, as well as spin-off companies that offer sensor integration and digital twin services.He added that the Augsburg AI production network has rented a 7,000-square-meter facility for this purpose, “expanding CosiMo’s development blueprint to a very broad scope, including linked automation cells, where industrial partners can Place machines, run projects and learn how to integrate new AI solutions.”
Carapappas said that Meggitt’s dielectric sensor demonstration at the NCC was just the first step in that.“Ultimately, I want to monitor my processes and workflows and feed them into our ERP system so I know ahead of time which components to manufacture, which people I need and which materials to order. Digital automation develops.”
Welcome to the online SourceBook, which corresponds to CompositesWorld’s annual print edition of the SourceBook Composites Industry Buyer’s Guide.
Spirit AeroSystems Implements Airbus Smart Design for A350 Center Fuselage and Front Spars in Kingston, NC
Post time: May-20-2022