key: cord-0756866-k8l1g2uf authors: Vanhooydonck, Andres; Van Goethem, Sander; Van Loon, Joren; Vandormael, Robin; Vleugels, Jochen; Peeters, Thomas; Smedts, Sam; Stokhuijzen, Drim; Van Camp, Marieke; Veelaert, Lore; Verlinden, Jouke; Verwulgen, Stijn; Watts, Regan title: Case study into the successful emergency production and certification of a filtering facepiece respirator for Belgian hospitals during the COVID-19 pandemic date: 2021-03-26 journal: J Manuf Syst DOI: 10.1016/j.jmsy.2021.03.016 sha: 03144218673b38bb5e9a38e6f37d5da355c8c50b doc_id: 756866 cord_uid: k8l1g2uf The SARS-CoV-2 pandemic presented European hospitals with chronic shortages of personal protective equipment (PPE) such as surgical masks and respirator masks. Demand outstripped the production capacity of certified European manufacturers of these devices. Hospitals perceived emergency local manufacturing of PPE as an approach to reduce dependence on foreign supply. The fact of a pandemic does not circumvent the hospital’s responsibility to provide appropriate protective equipment to their staff, so the emergency production needed to result in devices that were certified by testing agencies. This paper is a case study of the emergency manufacturing of respirator masks during the first month of the first wave of SARS-CoV-2 pandemic and is separated into two distinct phases. Phase A describes the three-panel folding facepiece respirator design, material sourcing, performance testing, and an analysis of the folding facepiece respirator assembly process. Phase B describes the redevelopment of individual steps in the assembly process The SARS-CoV-2 pandemic ("COVID-19 pandemic") requires personal protection precautions to be taken. These precautions are taken to firstly create a barrier against body fluids transmission between patients and health workers, and secondly as anti-viral respiratory protection for the health workers (through filtration). Mouth masks are one of the key products used as part of these precautionary measures. The COVID-19 pandemic has caused a worldwide scarcity of both surgical masks and filtering facepiece respirators ("FFRs"). for much flexibility or changes during the process. These traditional development process models are based on the idea that system requirements can be fully known at the outset. However, when there are vague or inadequate requirements, up-front planning appears irrational. It is common to have vague requirements at the start of a development process [2] . The traditional model is hard to apply during a crisis. This is because it is a slow and non-flexible model which is hard to implement with vague requirements and when the design specifications can change at any moment. When development speed and flexibility are crucial for a successful outcome, agile development appears to be a more useful process [3] . Agile development is a process that originates from the software development world [4] . It is an iterative, time-based, and result-oriented process, first introduced by Schwaber, K. in 1997. "Agile software development indicates software development methodologies based on the concept of iterative development, which creates requirements and solutions through cooperation among self-organized cross-functional teams." [5] One of the most commonly used design frameworks within the agile process is SCRUM. The SCRUM framework was developed in the early 1990s by K. Schwaber & J. Sutherland [6] . It is a lightweight framework that helps people, teams and organizations to generate value through adaptive solutions for complex problems. The agile development manifesto is made up of twelve principles that were described in 2001 by Beck et al [7] . Agile development was initially designed for software development, but it has also been applied in the product development context [7] . The differences between agile development for software and hardware systems was studied by Stelzmann [8] . Stelzmann analyzed existing work about agile software development and conducted interviews in systems development companies. He concluded that "in contrast to software, hardware systems that have to be produced physically often are difficult to be developed in small cyclic steps". He also stated that "Only if prototyping, testing, and implementing changes can be done quickly and cheaply, this principle is feasible'". Since the expiration of several key patents in additive manufacturing in 2010s, the prototyping industry has boomed and continues to grow exponentially [9] . Rapid prototyping continues to become cheaper and more accessible. This means that modern product development can overcome Stelzmann's concerns. It is therefore possible to apply agile development guidelines to the product development context. J o u r n a l P r e -p r o o f Natural and man-made disasters may occur at any time and are likely to occur more frequently due to climate change and environmental degradation [11] . In times of emergency, agile manufacturing and agile development have an important role to play to help society manage the consequences of natural and man-made disasters. An example of agile manufacturing and agile development can be found in 2005's Hurricane Katrina system that affected North America. At that time, Walmart and other large private enterprises wanted to protect their own physical assets. These large corporations therefore created their own in-house departments to plan for their recovery and response to natural disasters [3] . Walmart used agile manufacturing and development processes in their disaster planning and management. The easily expendable structure of Walmart's emergency response protocols "drives the ability to be agile and flexible" [12] . Walmart received universal praise for the way it responded to Hurricane Katrina [13] . In contrast, the United States federal government was far less effective at dealing with the natural disaster due to a lack of agility and resilience in its processes and applications. Many commentators have concluded that governments could learn from the agile processes applied by private corporations in order to prevent future failings in future disaster situations [14] . For manufacturing processes to be able to adapt during times of crisis they need to be smart and resilient. Linear waterfall processes have proven to fail when circumstances are uncertain and prone to change. In contrast, an agile process can help manufacturing processes to be flexible J o u r n a l P r e -p r o o f to changes during production. Some production techniques are better suited than others to cope with changing requirements. Additive manufacturing is a good example of a smart, agile process [15] . It can be used for small scale production where flexibility and lead time are a higher priority than production costs. These techniques are ideal if the process is prone to changes. Whenever certain requirements are set, the process can shift to more traditional manufacturing methods such as injection molding. Some other examples of agile manufacturing processes are laser cutting [16] and CNC-milling [17] . Many nimble manufacturing methods proved to be resourceful for immediate, local response to combat the corona virus. As shown by an in-depth case study [18] , several global efforts were established to mass produce ventilators, nasopharyngeal swabs and PPE such as face masks and face shields ( [19] [20] [21] [22] [23] . By using additive manufacturing, they were able to immediately meet urgent demand while overcoming challenges within the supply chain. Up-front prototyping describes a shift in how prototyping is used during the design process. In traditional design, prototyping is used only as a validation tool to evaluate the final design that has already been thoroughly analyzed [24] . Up-front prototyping describes a method where prototypes are made throughout several stages of the design process before complete analysis of the design. These prototypes can be used to test the design or its sub-systems by trial-and-error or at more This article uses this emergency production of filtering facepiece respirators as case study for agile product development. The four agile development stages were cyclically applied during the development process, with each cycle taking longer and becoming more detailed until a final design was crystallized. J o u r n a l P r e -p r o o f The development process can be split in two phases: Phase A and Phase B. In Phase A, an agile development methodology is used to create a certified, working mask design in the shortest amount of time possible. During this process, many iterations were tested until a working design was completed: The Minimum Viable Product ("MVP"). During this stage, the main priority is development speed and validation. There is a difference between developing something for mass production and developing something for a temporary local production line. In this section, the first design sprint will be described, which took place over a period of two weeks, beginning on 9 March 2020. The desired result of this first design sprint was a working prototype. Such 'proof of concept' would later allow the team to attract additional funding and government J o u r n a l P r e -p r o o f approval. In section 3.2, there is a description of how this prototype became the benchmark of an upscaled operation for a small production line, which eventually produced the first 100 respirators. After finalizing a working prototype, a second effort was made to improve the design of the mask with the generated knowledge, in order to adapt production to a larger scale. This time, a more traditional methodology could be used since the previous phase had resulted in sufficient production knowledge. The goal of this second stage was to manufacture 100 respirators per day, which exceeded the initial request from UZA for between 150-300 respirators per week. The priority in this stage had shifted to reliable and fast production at a larger scale, while still maintaining the quality performance aspects of the mask that would result in an effective product. In this second stage (Phase B) both the design and production environment were optimized for production at a small scale and future upscaling possibilities. The most crucial element of a respirator mask is the filter material, this material must be able to filter the virus. The material must be tested according to EN149. PTFE nanofilter media for FFP2 or FFP3 classes. FFP2 media was in A4 sheet format. FFP2 media was on a roll On the outside there should be a layer that protects the user from fluid spatter (blood, coughing). Polypropylene sheet (PP), 50 grams. On the inside of the mask there should be a layer against the skin for comfort and moisture absorption. A new valve was designed and patented. It was designed specifically for 3D printing and rapid production. The first step of product development is reverse engineering and morphological mapping. This is generally done by creating a map that J o u r n a l P r e -p r o o f includes every sub-part of the product, generating a multitude of design ideas for each sub-system and subsequently selecting and matching the best design ideas. The result of such a process is illustrated in Figure 1 and Table 3 . A fast way of gathering information about a certain product is to look at existing or similar products on the market. The process of analyzing a product to learn about its design is called reverse engineering [32] . The team reverse-engineered an commercially-available three panel, flatfolding FFR commonly used by local hospitals, to quickly gain enough knowledge about FFRs to kick-start the development of an in-house model. Reverse-engineering enabled information to be gathered quickly about size, materials, production process, assembly, performance and other factors. One of the first things to become clear during the reverse-engineering process was that a respirator's nose bridge, foam strips and head harness are crucial components for ensuring a good face fit, which is essential for a proper seal. When observing existing respirator designs, it became apparent that these elements are easily underestimated and swapped for cheap sub-quality materials. However, these three components need to be thoroughly researched and precisely matched to create a functional respirator. It was challenging to locally source good materials for these elements, but without a good face fit an FFP2/3 respirator is useless. J o u r n a l P r e -p r o o f The sourcing of materials was a big roadblock to starting up a production line during a global pandemic. A worldwide supply chain disruption caused shortages. This deficiency is being caused by an exponentially growing demand for PPE, panic buying of common items and the temporary shutdown of factories and logistics companies. Sourcing was challenging due to a high dependency on international suppliers. Due to regional lockdowns and the closure of factories, suppliers and logistics companies, as well as governments requisitioning stocks, resources became hard to find internationally. It therefore became clear that local suppliers had to be identified so that materials could be accessed quickly. However due to globalization, many of these supply companies had moved to cheaper countries such as China [33] . Further, even when materials were offered by lesser-known international suppliers, their quality was questionable. A lack of Europe certification prevented the Antwerp Design Factory team from using these materials in our prototyping. To enable prototyping to begin as swiftly as possible, small quantities of sample materials from local suppliers were used to make first proof of concept prototypes. Those limited quantities of materials were easy to obtain while waiting for larger batches of materials to arrive. Examples of materials our team sourced are shown in the chronological development timeline in Figure 1 . The key material for a filtering facepiece respirator is arguably the filtering material that stops viral penetration. Most commonly, melt blown fibers are used for the fabrication of filtering face pieces. It is a one-step production process that produces self-bonded fibrous nonwoven membranes directly from polymer resins, with an average fiber diameter ranging between 1 and 2 µm. This material has proven to be highly effective at filtering nano sized particles [34] . The fibers are then electrostatically charged using a corona treatment process to achieve a higher filtration without increasing the pressure drop [35] . However, this manufacturing process is difficult and costly to set up. As a result, stocks became quickly depleted, and global demand was higher than the global production volume. This resulted in material prices increasing by up to 25 times the normal price [36] . Since the pore size and electrostatic charge of feasible filter materials must be known to evaluate the effectiveness of filtration [37] , only few potential alternative materials were available. Since the team did not have the resources and infrastructure to test the actual filtration performance of materials, fabrics with already known certified filtration properties were selected. One of the most commonly-used filtration rating systems are HEPA and MERV [38] . Therefore, it is important to cover this part of the mask with a soft foamlike structure to seal all possible gaps between the nose and the mask. Sourcing this material proved to be harder than anticipated. To achieve a good seal, the material must have several specific properties such as the right hardness, a closed cell structure to stop leaks, biocompatibility, non-toxic, latex-free, easy to cut and have one side covered by pressure Figure 2 shows the results of quantitative fit testing of the team's prototype using a 1.5mm-thick foam with 10A Shore hardness from VMX Silicon and comparing this to 3.2 mm-thick Polycushion Padding Sheet from Rolyan. The respirator prototypes were tested three times using the N99 test The aluminum nose bridge was supplied by a local arts and crafts store. Several colors were validated and were reported to have diverse gluing J o u r n a l P r e -p r o o f performance depending on the selected surface treatment. Non-woven PP sheets both 30g/m² and 50g/m² were sourced with relative ease from a local company (Voltex NV, Torhout Belgium). Elastic straps were also easy to source, however after evaluating several types it became clear that they needed to meet several requirements. The straps had to be latex free, which highly limited the available choice. Another essential factor was a high elasticity and an elongation factor of at least 300%. High elasticity is needed to ensure a tight seal for a range of different head sizes, whilst not pulling too hard on larger heads leading to discomfort. It was eventually possible to find an elastic strap material that would be adequate for a minimum viable product. As noted above, filtering facepiece respirator certification is a challenging process. However, rapid in-house validation methods had to be identified to objectively validate the quality of the protypes during the design process. For this purpose, the team selected three tools: 3M™ Meter to build a breathing resistance instrument around. The FT-30 is a test kit that includes a see-through hood that can be placed over the shoulder of the test subject wearing a mask. A nebulizer is used to disperse aerosols of a mixture of water, sodium chloride and J o u r n a l P r e -p r o o f denatonium benzoate into the hood. A series of standardized tests are performed by the user. This method is used to verify the mask to face seal. When the seal is not completely closed, the user will taste the bitter aerosols which verifies a leak. This method is used by hospitals around the world to verify mask fit before entering a contaminated environment. This test was used initially by our team to quickly verify the mask fit during the early design process. However, this test kit is unsuitable for subtle refinements in mask design. This method also cannot be used to test the filtration properties of the mask. A switch was therefore made to a more trustworthy method of testing using a quantitative tool: The PortaCount Pro+ Respirator Fit Tester 8038 (TSI Inc. MN, USA). This quickly became the most important machine for testing respirators. As described by [40] the PortaCount compares the number of aerosols inside the respirator (Cin) with the number outside the respirator (Cout). A fit factor is generated that is equal to the ratio of these two measurements (Cout/Cin). To comply with the OSHA requirements a fit factor of 100 should be achieved for N95 respirators. This means that the air inside the respirator is 100 times cleaner than the air outside the respirator, within the measurement range. More specifically, the PortaCount 3038 measures a concentration range from 0.01 to 5×10 5 particles/cm 3 with particle sizes from 0.02 µm to greater than 1 µm [41] . The fit factor is determined by the total inward leakage. This leakage is the sum of the particles penetrated through the filter material and through holes due to insufficient fitting to the face. When the filter media is certified and complies to the desired filtering requirements, the PortaCount can be used to analyze the fit of newly generated prototype FFRs. It is important to understand the breathability of the components that This setup was used to measure the breathing resistance of commercially-available masks and samples of filter media that our team would source. An example of this is shown in Figure 4 , where the back pressure was measured at high and low flow pump rates (5.5 L/min and 2 L/min, respectively) for a standard Type-I surgical mask, a 3M Aura 9320+ FFP2 respirator, and a 3M Aura 1883+ FFP3 respirator. Samples of the FFP2 and FFP3-class PTFE nanofilter materials were also measured. Importantly, Figure 4 shows that at both low and high flow rates, the melt-blown filter media in the two 3M respirators are more breathable than the PTFE porous membrane used in our team's prototypes. Due to the high filtration efficiency of the FFP3-class PTFE nanofilter, breathability issues were likely to present themselves compared to conventional melt-blown polypropylene filter media. The way commercial FFP3 facemasks work around this is to install a one-way valve to assist the exhalation of the wearer. A novel design was made for an exhalation valve that is optimized specifically for 3D printing and local assembly, which comprises of three 3D-printed rigid valve bodies, a flexible diaphragm to create the seal made from a latex material, which is secured to the filter material of the respirator using an M3 stainless J o u r n a l P r e -p r o o f The prototyping phase was a matter of reverse engineering and fast failing. By using an agile development methodology, the team was able to cycle through all phases of the product development process almost every day. This resulted in a new prototype being produced on a daily basis. These mock-ups were mostly made up of free materials samples from local companies, 3D printed valves, and stapled elastics. In Figure 7 , an overview of filtering facepiece respirator prototypes is shown, numbered sequentially from 1 to 16. The 16th prototype was the team's Minimum Viable Product (MVP). All these prototypes were produced during the first two weeks. A total of 6 steps were required to create a single mask, with a combined manufacturing time of 35.5 minutes per mask. All these steps could be performed manually by a single person. This is summarized in Table 4 . It was then concluded that a second design stage was necessary to optimize the production time and ensure the product could be manufactured at a small industrial scale. J o u r n a l P r e -p r o o f In the second phase (Phase B), a small-scale production line was developed that could make more than 100 masks per day. An overview of the production steps is shown in Figure 9 . Step 3a is an ultrasonic sewing machine, Step 3b is semiautonomous ultrasonic point welding using a Cobot, and Step 4 is a manual ultrasonic cutting station. In order to further optimize the production of FFRs, some adjustments were necessary. A first change was mounting the rolls of filter material and spun-bound polypropylene materials on a fixed gantry frame, which then required less preparation time per respirator. In addition, the team gained access to a clean new laser cutter that would help minimize contamination. Next, an ultrasonic sewing machine was made and put into use for welding and cutting the seams of the upper and lower panels. In this way, much more consistency in the seams of the mouth mask was J o u r n a l P r e -p r o o f achieved. It also significantly reduced the time needed for manual ultrasonic point welding. Furthermore, a cobot was programmed to weld the three different panels together. This required a small adjustment to the shape of the filtering facepiece respirator which made it possible to fix it in a predetermined position underneath the cobot. In this way, the cobot arm could weld the panels together considerably faster and more consistently than a single person. Each of these changes is described in more detail in the following sections. The transition from making a few masks per day to a 100 masks per The starting point to determine the new respirator shape was a threepanel, flat-folding FFR used by UZA. These FFRs were the preferred product of the healthcare workers but were no longer available due to supply chain failures caused by the COVID-19 pandemic. The flat-J o u r n a l P r e -p r o o f folding FFR's shape was scanned and converted into digital drawings. The three-panel design was prototyped into a functioning respirator using the techniques explained further in this section. Using the only available anvil for the sonotrode resulted in wider welding seams compared to the commercially available design. A final adaptation was made by adding outside clamping points to the straight edges of the shape, to fix all three panels in a magnetic clamping jig during the cobotic welding. The final design is shown in Figure 10 . To improve the welding quality of the chin and nose panels, the manual welding method was replaced by an ultrasonic sewing machine that was equipped with a titanium rotating anvil with a toothed-wheel and the cutting blade, as shown in Figure 11 . A drive system was added for the anvil to enable automatic feeding of the material. This driving system was linked with the activation pedal of the ultrasonic welder. This configuration meant that the functionality of this machine was comparable to an industrial ultrasonic sewing machine. In order to improve and speed up the welding of the three panels, a Franka Emika Panda 7-axis cobot (Franka Emika GmbH, Munich, Germany) was introduced to the production. A Cartesian coordinate system was used to define the position of the welding points, show in For the nose bridge, an annealed flat aluminum square wire (1.5 × 3.5 mm) was selected and cut to length by hand with side cutters. Given the simplicity of this process, automation was not necessary to improve production speed. Therefore, the same method was used throughout the whole process, allowing for flexibility. To attach the nose bridge to the mask, a high-end hobby hot glue gun (Steinel Gleumatic 5000, Steinel Group, Clarholz, Germany) was used. After gluing, a palette knife was used to apply even pressure during the solidification of the glue. To prevent the nose bridge from peeling off, an extra dot of glue was applied on top of both sides of the aluminum strip. The increased filtration of FFP3 masks results in a higher pressure drop, which correspondingly makes the mask harder to breathe through. To eliminate sub-optimal breathability and moisture and CO2 buildup, a one-way exhalation valve is required. An exhalation valve was 3Dprinted in three parts, while an M3 nut and bolt were used to assemble all parts and fix a circular disc into place which acts as the valve membrane diaphragm. The membrane was punched out of stacks of the same elastic material as the head straps. The requirements for CPA evaluated by IFA is that the respirator needs a name, that the respirator has labelled packaging, and that a manual for usage is included. Our team chose to call the respirator the ADF3 (Antwerp Design Factory FFP3), and respirators were packaged in ziplock bags with printed adhesive labels attached to indicate product code details. Finally, usage instructions were created that described donning and doffing procedures and disposal of the respirators. J o u r n a l P r e -p r o o f A previous study concluded that the flexibility of many manufacturing methods has difficulties adapting to changes since they are constrained within a very limited boundary. Robots, particularly cobots, in combination with modern manufacturing technologies can prove useful to ramp up production in emergency situations. This was shown in a case study about emergency ventilator production [44] . To enable the team to make the production process faster and more consistent, a Franka Emika Cobot was used. This resilient manufacturing robot allows for easy programming and configuration. Combined with a clamping jig, the three panels could be aligned and welded accurately and consistently using a custom program. To achieve consistent pressure, a 3D-printed end-effector for the sonotrode was designed with a built-in spring system, as shown in Figure 13 . This allowed the sonotrode (which was initially developed to be used for cutting) to be mounted at an exact angle and weld with a consistent pressure. A total of three jigs were used so the production line user can replace the welded mask with new panels, A small assembly of six people was arranged to sequentially assemble the individual components of the FFRs. Using this approach, more than 100 respirators were able to be fabricated in an afternoon, following the steps outlined in Figure 9 . The interaction between the stations can be seen in this video summary: Product developers produce first 100 masks (april 2020). In Table 5 , an overview of the improved production is shown. It shows a total production time of 553 seconds per mask during Phase B, of which nearly 200 seconds were material preparation steps. In Phase B the team used six people in the production line. This resulted in 120 masks being produced in 4 hours, which equates to 30 FFRs per hour. This was a significant improvement from Phase A, which required 35.5 minutes per respirator. Even if six people were used in Phase A to simultaneously perform the individual preparation and production steps, the slowest station, manual ultrasonic point welding which took 12 minutes, would limit the yield to only 5 FFRs per hour. Photographs of the fabricated masks first are shown in Figure 14 . Once these certifications were in place, the remaining 85 masks were transported to UZA for their use. A short summary of this delivery is shown here: First delivery of respirators to UZA. The COVID-19 pandemic created a unique set of circumstances: a major Belgian university hospital requested that a university research group establish an emergency production line of respirator masks. This was to ease chronic PPE shortages in European hospitals due to global supply chain failures. The Antwerp Design Factory answered this call for help and designed and fabricated respirator masks that fully met the hospital's expectations and European safety regulations. Apart from a broad objective to produce between 150-300 FFP2 or FFP3 respirators per week for UZA, the team was given little input or guidance. An agile product development process was therefore implemented to overcome uncertainties. The team's efforts are a successful case study of resilient and smart manufacturing in the COVID-19 pandemic. There were two distinct phases in this project: Phase A and Phase B. Phase A consisted of a sprint towards a minimum viable product which satisfied the team and the hospital's requirements for performance and comfort. During Phase A, the sprint approach required team members to take responsibility for different aspects of the respirator development. process revealed that the manual ultrasonic welding would limit production yield. In addition, the manual spot welding is a very tiring process. This is because a hand-held ultrasonic welder is awkward to control and tiring to use for extended periods, due to factors such as the weight of the hand-held pen, the driving cable and the air-cooling line. Phase B refers to the team's upscaling strategy to move from the MVP to the reliable production of 100 respirator masks. An important decision in this phase was to utilize a cobot to perform the repetitive ultrasonic spot-welding tasks. The use of the cobot in this application is an example of innovative adaptation of human-machine interface. The cobot added J o u r n a l P r e -p r o o f resilience to the team's production line, as it removed human error (for example caused by fatigue) from the fabrication process. An important secondary contribution was the development of a novel spring-loaded end-effector for the cobot to hold the pen-shaped ultrasonic welder. Again, 3D printing was employed as a smart manufacturing solution which allowed the design of the end-effector to be rapidly iterated. This meant that the cobot could apply spot welds that would mimic the feel or touch that a human operator could achieve. Ultimately, 120 respirator masks were manufactured in a single afternoon using the Phase B production plan outlined. Of these masks, 20 were sent to IFA for the German CPA testing and 15 were sent to Mensura for the Belgian ATP. The ADF3 respirator passed both these certification procedures, and the remaining ADF3 respirators were shipped to UZA. The hospital could therefore be certain that the Antwerp Design Factory's respirators fulfilled all relevant European and Belgian regulatory requirements. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. The following statements describe the contribution of each Author: Product Design and Development: Fifth Edition. Fifth edit Agile Product Development: Managing Development Flexibility in Uncertain Environments Wal-Mart to the rescue private enterprise's response to Hurricane Katrina The Art of Agile Development. First edit Digital Template Market. Why Agile is important for Software Development -Digital Template Market Manifesto for Agile Software Development Twelve Principles of Agile Software Contextualizing agile systems engineering The evolution of rapid production: How to adopt novel manufacturing technology Design of agile supply chain assessment model and its case study in an Indian automotive components manufacturing organization The impact of climate change on natural disasters How Wal-Mart Beat the Feds to New Orleans Best Responders: Post-Katrina Innovation and Improvisation by Wal-Mart and the U Executive Office of the President, et al. The federal response to Hurricane Katrina: Lessons learned. Government Printing Office Additive Manufacturing and PSS: a Solution Life-Cycle Perspective Hybrid lean-agile manufacturing system technical facet, in automotive sector Integration of rapid prototyping technology into FMS for agile manufacturing Additive manufacturing and the COVID-19 challenges: An in-depth study 3D Printing of Face Shields During COVID-19 Pandemic: A Technical Note From making cars to ventilators 3D Printed COVID-19 Test Swabs Fast Radius. Request reusable medical face shield kits 3D-Printed Non-Invasive PEEP Masks Aim to Alleviate Ventilator Shortage Adapting scrum development method for the development of cyber-physical systems Regulation (EU) 2016/425 of the European Parliament and of the Council of 9 March 2016 on personal protective equipment and repealing Council Directive 89/686/EEC Commission Recommendation (EU) 2020/403 of 13 March 2020 on conformity assessment and market surveillance procedures within the context of the COVID-19 threat on Conditions with which the supply of mouth masks FFP2 and CIIRC CTU Develops Own Prototype of CIIRC RP95 Respirator / Half Mask 2020 Additive manufacturing for COVID-19: Devices, materials, prospects, and challenges Coronavirus: non-conforming respirator masks -Alternative Test Protocol Federal Agency for Medicines and Health Products Reverse Engineering and Design Recovery: A Taxonomy China's embrace of globalization. No. w12373 Fabrication of nanofiber meltblown membranes and their filtration properties Method of making fibrous electrets Meltblown price evolution in China from Melt-blown filter medium HVAC Filter Selection and MERV Ratings: What Does It All Mean? Ratings Scale and Filter Efficiency Fit Testing Healthcare Professionals Can Trust Portacount R Pro 8030 and Portacount R Pro+ 8038 Respirator Fit Testers Operation And ServiceManual Role of additive manufacturing in medical application COVID-19 scenario: India case study Additively manufactured respirators: quantifying particle transmission and identifying system-level challenges for improving filtration efficiency Reconfiguring and ramping-up We would like to extend our gratitude to the KU Leuven Robotics The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.