Toward the end of 2019 LayupRITE started its follow-on Ufi funded project. The main aims of this project, titled “LayupRITE101: Augmented Training for Composites”, was intended to integrate the LayupRITE system into a more complete training course. For this project, LayupRITE would be used to augment the teaching of the NCC’s “Introduction to Manual Prepreg Layup” course. This 2-day course gives both theoretical and hands-on instruction in the manufacture of prepreg components by manual layup. This includes safety, storage, materials, tooling, layup, consolidation, and finishing operations. Along with an e-learning component designed to enhance the current classroom-based course notes, LayupRITE simulations would be used to demonstrate theory points and guide the workshop tasks.

Feedback from Ufi Project 1

The first Ufi-funded project gave the LayupRITE developers a lot of useful feedback and learning points. Some of the points related to “what good looks like” for augmented reality backed learning, others related to cost and form-factor concerns.

In terms of “what good looks like” that was missing from the first project, the key element was in being able to make mistakes for practice. The drape algorithm solved the drape for the entire net, meaning that the user could only really change the start point. This was still useful as the start point in a layup generally determines the final ply outline and was taken as a learning accelerator. Project 1 also demonstrated that, without guidance, people will use a variety of starting locations.

Feedback from our target customers, further education institutes, was that the projector-based system would either be too costly or too bulky for the environments they would use it in. Additionally, during development there were issues with setup time and calibration, as well calibrating between setups. As a result of these concerns, the form-factor and AR delivery for LayupRITE101 was chosen as tablet-based, pass-through AR. The intention from the developer was that the same computer could be used in both the classroom and the workshop to guide the exercise.

LayupRITE101 Drape Simulation

“Facilitating deliberate practice”

The Pin-Jointed Net method for a kinematic drape solves the drape for the entire net and reports back the resulting net shape and shear angle. Previous work [1] has shown that drape simulation doesn’t necessarily match the feature-by-feature approach employed by laminators. Additionally, whilst it quickly solves the drape of a virtual net with next to no initial input, save for a start point in our case, it doesn’t allow the user to introduce (deliberately or otherwise) defects. The method also lacks the interactivity which benefits learning.

The pin-jointed net assumption does lend itself to quickly trialling start points. Start point is a key accelerator in learning how composite plies drape. As such, it was made into its standalone lesson “module”. The 4 lesson modules developed for the course will be discussed later.

What was useful for this project was the development of a simulation that could show how the net should deform in real time. What would be even more beneficial would be a net that could be interacted with, so a user could practice draping a virtual net. This virtual net would differ from the automatically solving net in that the user would be able to interact with it and attempt to drape it in the more natural feature-by-feature style. This would require creating a new type of manually controlled virtual net in the Unity environment.

What this manual net gave us was a net where nodes could be fixed in space or moved. What this enabled was in pre-shaping or folding the virtual net prior to placing it on the tool.

 

By fixing some nodes and moving others the net was also able to demonstrate another key learning point for composite drape: bias-extension shear. Fibres are strong, that’s the point, so pulling in the fibre direction won’t allow you to deform, shear and shape the net. Pulling in the direction off the fibre axis (the “bias” direction) is how a laminator can shear the net to make it conform to the mould tool. This was also a key learning accelerator and was made into its own module.

Drape Simulation Modules

In total 4 modules were developed and trialled within LayupRITE101In addition to the two accelerators mentioned earlier, start point and bias extension, there was a “free practice” module and the workshop exercise.

Bias-Extension and Shear

The bias-extension less is a quick (under 5 minute) exercise to demonstrate the learning accelerator of the same name. The final version is compatible with both mouse and touch controls and shows the shear increasing by the net changing colour and eventually displaying an “X”-shaped symbol when the cell is over-sheared. The exercise demonstrates pulling the edges and corners on both 0°/90° and ±45° nets.

Start Point

This exercise uses the automatically draping nets to compare starting points. The user uses the mouse or touch to select the starting node and tap/click to begin the drape. There is a text readout of some net statistics (net fibre angle, max shear, average shear, % over-shear) which compares the last two nets positioned. Additionally, the graphic will show the outlines of both nets and the full shear mesh of the topmost net.

Workshop Exercise

This module uses script controlled manual nets to give step-by-step instructions for the workshop exercise. This includes ply order, orientation, location, location of any joints/overlaps, material type, and cores/adhesives etc. The module also has an exploded view for easy visualisation and is intended to effectively replace the paper ply book and work instruction. The module also included an AR function where the virtual plies would line up relative to printed AR markers positioned on the tool. Feedback from users was very positive about the “virtual ply book” aspect of the module and less positive about the AR aspects/implementation. However, this module gave confidence in the whole system and possible expansion routes for LayupRITE.

VFP Basic – Free Practice

Unlike the other modules, VFP Basic doesn’t have specific guidance and is intended as a practice area for users. Tools (in *.stl format) could be loaded in and automatically sized nets in any fibre orientation could be created. This includes both automatically and manually draped nets and their shear meshes would be displayed visually in the program. Nodes can be fixed in space, selected, and moved or left as “normal” with their movement/position to be solved by the simulation. When a node contacts the tool surface it sticks and becomes “draped”. To lift the stuck node off the tool there is an “undo” tool. This is typically more challenging in the real world, but it allows users to make and fix mistakes, important for learning.

Trials

There were two sets of in-person trials planned, at the National Composites Centre and with partner further education institutions. Unfortunately, due to the global Coronavirus pandemic, these plans were severely curtailed. The NCC trials were still able to go ahead with masks and physical distancing, but the offsite trials had to be made virtual. This necessitated some additional software management, as the original plan had been to have the programs on NCC hardware, which the trial team would take away the end of the trial.

Respondents were asked to complete a survey and the LayupRITE system was generally well received. The top-level score out of 5 are summarised here:

Modules	Rating (Out of 5)	Responses
Bias-Extension Lesson	3.67	12
Start Point Lesson	4.17	13
Workshop Session	4.43	7
VFP Exercise	3.80	10

The workshop exercise received the most favourable feedback, with the VFP practice session ranked the lowest. The common response for this was that the VFP exercise lacked the proper context within the trial session. For most the sound design was also a point of negative feedback, with the quick fix being the inclusion of a mute button.

The AR implementation was also reviewed negatively and requires further development for its accuracy and ease-of-use. This was expected by the team given some hardware and development issues and is also seen as a fixable issue, given the proliferation of similar AR applications.

LayupRITE101 Conclusions

This project has allowed for great additional development for the LayupRITE simulation system and learning toward requirements for integrating into an existing course. The flexible software design and e-learning content mean that information can be easily extracted and repackaged into other courses, where relevant. The use of a digital ply book with an exploded view allowed the instructors to spend less time demonstrating and more time interacting with learners, this was a great outcome for the reasonably limited trials (due to the pandemic).

There is great potential for further developing the exercises and accelerators as standalone learning elements as well as further software improvements. The workshop exercise/virtual ply book has the potential to be developed into its own work instruction generation methodology and is a subject of further work and development.

At this point we would like to thank the Ufi Charitable Trust for funding and supporting this project as well as the National Composites Centre for facilities and trials help.

This series of posts is intended to showcase the top-level outcomes of Ufi Project 1 titled “Augmented Learning for High Dexterity Manufacture”. This project was funded by Ufi, a vocational learning charity. In this post we will be briefly looking over the outcomes of user trails. As part of the Ufi project trials were conducted with the NCC, two colleges and an SME. Since everyone outside of the NCC was volunteering their time, we sought to make the trials as unintrusive as possible, whilst still giving us useful information to work with.

Table 1: Total trial participants

Location Type	Number
College	31
NCC	23
SME	8
TOTAL	62

SME and College Trials

The SME and college trials were used to give us feedback on the “form and function” as well as suggest use cases for the system. It also enabled us to see how the system worked with example customers and get their thoughts. The system overall was generally well-received by students and laminators however, some did note the lengthy setup, and felt that the system needed further development.

In the college setting a key point of feedback was the unit cost. At the time of the trials the cost of the PC, projector and Kinect was somewhere in the region of £2000, the projector being the bulk of the cost. This did not include software. Some accommodations could be made, in that any Windows PC could possibly be used, although there are some compatibility issues with the Kinect and certain brands of USB cards. Another suggested solution was that one system be used with multiple users. This was certainly a possibility given that the Kinect has the capability to track and identify up to 6 users simultaneously.

Trials with the National Composites Centre (NCC)

No Instructions

Laying-up the ply with no instructions gave us confirmation of a widely held belief in manual layup: everyone does it differently. Using video analysis, we determined the first point of pressure of the ply onto the tool, this would be the modelling equivalent of the seed point. There was a bias toward the middle of the component, or at least the centreline, which can be attributed to the part being symmetrical, however there is still a clear spread.

There was also one instance of a laminator cutting a ply on the tool to ensure proper fit. This is a relatively common occurrence with manual layup and the laminators were asked to use their best judgement about what to do to lay up the ply. Unfortunately in this case, the laminator actually caused damage to the tooling which might have effected the ability to de-mould the finished component.

These kinds of events could potentially go unrecorded under normal circumstances. Generally, the quality was somewhat mixed, due to time constraints and some material issues, so LayupRITE was unlikely to be the decisive factor in quality comparisons.

The sets of images above show the different routes laminators took to lay up the part. In addition to different starting points their processes were both different. In these selected examples, the laminators are both going back on themselves, rather than working out from the start point, as drape models would suggest is optimal.

Using LayupRITE

The LayupRITE instructions for this component were developed in a previous project with their aim being to symmetrically drape this component from a seed point. Since the next step in the process won’t appear until the current stage is complete, the participants all followed the projected instructions. This means that all the individual plies were laid-up using the same route. This route was not actively optimised but was intended to drive consistency. Additionally, by following the prescribed pattern, cutting the ply on the tool was not needed, so no “quality incidents” like the one mentioned above occurred.

Timing Results

Table 2: Comparison of time to lay down ply average and range

	No Instructions	Using LayupRITE	% change
Average	6m 30s	5m 53s	-9.6
Range	7m 40s	6m 4s	-20.9

 

The above table shows that using LayupRITE instructions reduced the average time to drape the ply over the U-shape tool. What is important to note is that it actually made the quickest laminator slower (a highly experienced technician), but that the average was reduced across all participants. The range of times was also reduced, this is important for planning out jobs since you know given the instructions roughly how long the steps will take.

Outcome of Trials

The trials showed us that the prototype system was useable but needed some further development before it could be a saleable product. Its main strengths were in that it was easy to understand and use, everyone “bought into” the idea. This is a key positive as it shows that we were along the right lines. The timing data showed that the system could be used to improve standardisation as well.

In terms of drawbacks the most obvious is in the setup requirements. The physical setup and calibration are both fairly lengthy processes but can be streamlined through further development. The physical setup can also be improved if a fixed, dedicated workstation is used. A secondary aspect which relates to the system was the cost for FE colleges. Whilst this wasn’t necessarily an issue for SMEs and larger companies, a branch off this system which is lower-cost would be useful to serve those potential customers.

This series of posts is intended to showcase the top-level outcomes of Ufi Project 1 titled “Augmented Learning for High Dexterity Manufacture”. This project was funded by Ufi, a vocational learning charity. The main difference between this and previous works was that the focus was on skills training. Training had always been touted as an application for LayupRITE, but this was the first time it was the specific goal. This gave the project two opportunities: firstly, to further develop the LayupRITE system and secondly to get a closer look at training as an application.

Skills training and Horizon Scanning

A key difference in was required when thinking about skills training. Previously, our strong suit had been in drape simulation and working toward making unambiguous instruction sets. Going into this project we believed that we could make a series of moulds of increasing complexity and walk the learners through. However, it was explained to us that that wouldn’t necessarily do any thing for the retention of the information. This was the best explanation we were able to come up with:

IKEA do great instructions, but if you were to take those instructions away, would you be able to assemble that wardrobe tomorrow, or next week? Would you know how to assemble a similar, but different wardrobe?

We also had to understand what good looked like from a learning design standpoint and what was out there already. To achieve this, we undertook a “Horizon Scan” of the current landscape of skills training in composites, current augmented reality (AR) applications and a study of learning theory and instructional design. The top-level outcomes from these three pillars were:

The composites training and AR applications pillars gave us encouragement that there was a space for LayupRITE to exist. There were a variety of AR applications in other industries and there appeared to be an opportunity to modernise, digitise and “smart”-ise composite laminator training.

A particularly interesting application was Soldamatic. Their system uses a welding visor/headset and torches with AR markers to better simulate the working environment. The system displays the material type overlaid onto real-world models of components to be welded. What is of interest here is how it ties in with the learning design findings, particularly “fading feedback” and “Cognitive – Associative – Autonomous”. During the course of the Soldamatic system the heads-up display in the visor displays less and less information as the user gets more experienced. This is a great example of “fading feedback” and ties in with the “Cognitive – Associative – Autonomous” approach to learning.

The Cognitive – Associative – Autonomous Model

• Cognitive – The learner is being told what to do and must think about how to do the task
• Associative – The learner understands what to do and can predict outcomes
• Autonomous – The task is performed instinctively, the focus is on strategy and efficiency

7 Principles for “What good looks like”

Finally, the pillars of the Horizon Scan led us to 7 principles for what good looks like:

1. Learning outcomes and performance standards to be achieved are clearly defined

• the precise and detailed analysis of the skills and processes, and the range and degree of difficulty of these to be covered
• the accuracy, speed and consistency with which they need to be undertaken
• the expected capability to transfer and adapt their application to different circumstances

2. The learning programme takes account of the stages of skills acquisition and the level of expertise of the learner

• starting with prescriptive guidance of generic skills, moving toward the information received in a manufacturing context
• fading feedback as learner moves through programme (and transitions through skills acquisition stages)
• self-direction and autonomy of learning programme (will need trainer intervention/assessment as well)

3. A low risk, low-cost environment provides for relevant and deliberate practice

• low risk – training environment lower risk than in-house training on real parts
• low cost – attempt to simulate material (material is largest cost in training), lower cost than taking experienced staff away from production, aim to accelerate skill acquisition
• deliberate practice – user control of programme

4. Guidance and feedback is targeted on what the learner needs to accelerate their skills acquisition and presented as simply as possible

• “as realistic as necessary” – animations preferred over video
• “multi-modal” – explore options for audio as well as visual feedback
• multi-screen – use of secondary display for more detailed/ancillary information

5. The learning programme enhances and supplements, if necessary, the intrinsic motivation of adult learners

• evidence of competence displayed to the learner (some gamification) – mastery
• show where these skills are used (e.g., high performance auto/aero parts) – esteem/recognition
• user-control over learning programme – autonomy

6. Evaluation is built-in from the outset and enables continuous iteration and improvement

• Relevance – relevant evidence provided
• Facilitation – development of effective accelerators
• Transferability – does the training transfer into the real world?

7. Attention is given to the whole learning context, not just the technology

• Practical issues – how the tool is set up and used
• Learner perspectives – introduction to tool at different levels of experience etc.
• Trainers and coaches – roles, support learners, how the tool supports them
• Wider environment – employers, awarding bodies, product design, quality control etc.