Page Index

Executive Summary

The Covid-19 Pandemic has highlighted challenges in getting immediate access to ventilators or resuscitators especially for patients needing to be transported between medical facilities while experiencing hypoxia. We learnt from our Stakeholders, doctors in Gujarat, India that in such scenarios the patient is manually ventilated using a BVM by ambulance personnel. These medical personnel can often be understaffed and overworked. Pumping a BVM for journeys of upto an hour pulls them away from other aspects of critical patient care, increasing fatigue and chances of human error.

We as a team decided to tackle this challenge by designing an automated BVM pumping mechanism that is cost-effective, compact, lightweight and robust. As strong advocates for echo-sustainability we were motivated to devise a solution which could reuse a lot of existing technology. Hence, the design showcased below will be an automated version of pre-existing and readily available medical resources with better assistive and failure detection features, making it a creative bio-tech solution.

Initially, leading up to the MVP after rigorous research and discussions with our Stakeholders we decided to attempt a solution which was similar to the ones tested in the industry, aka a motor driven pumping mechanism. We implemented this by refurbishing an unused 3D printer but we realized that the structure was quite bulky, intrusive, more suitable for bedside use, and was unable to compress the BVM sufficiently. We also found that most of the other potential solutions were built around this mechanism. So for our final product, we wanted the design to add minimal structural overhead and spatial inconvenience to the paramedics, patients and accompanying kin.

On exploring different design avenues, keeping our requirements in mind, we deduced that the Pneumatic Design was the best fit for our use case. We consulted with different industry experts who suggested Festo as the leading manufactures in the field. This design choice helped us make a unique, eco-friendly, portable and compact design such that it could be easily embedded in an ambulance environment.

So what does our Product Do?

Our product as-is meets all the criterias that we set out to achieve, they were decided on the basis of extensive research and stakeholder feedback. While remodeling our MVP design the following goals and requirements were concluded upon for the final product:

Design Architecture

Pneumatic System-as-Built:

Mechanical Diagram of the Pneumatic Pumping mechanism

Mechanical Schematic of the Pneumatic Pumping mechanism

Block diagram of the Pneumatic Pumping mechanism

Electrical and Software System-as-Built:

Schematic diagram of the Electrical Components of the Pneumatic Pumping mechanism

Electrical diagram of the Pneumatic Pumping mechanism

Code

Performance Analysis

For our product to work as desired, the pneumatic components that draw power should be connected as mentioned above and powered on correctly. Once the circuits are made and pressurized air is provided, the piston will pump at 14 breaths/min. We ran several tests by modifying the periodic switching between 24-0V, successfully varying the pump rate between 12-20 breaths per minute. The duration of time the system is left running has little effect on the pump rate, deviation being ±1 breath/min. The 14 breaths/min were achieved by testing out different parameters for the time.sleep() function which is called twice, once during extension and again during retraction of the pneumatic cylinder piston, as seen in the code (images/diagrams/code.PNG)

The system is left running for about an hour to check for overheating and clogging, there were no anomalies in operation recorded. We did however note that a compressor that can pump upto 80psi would be sufficient as opposed to the max. 120psi one which has a large reservoir tank. The air compressor we connected to for testing had a holding capacity of 11.1L which proved to be quite small, leading to the compressor being turned on rather frequently(every 30 seconds). We noticed that with a pressure setting of 50psi, we were able to squeeze the bag a desirable amount but the pump rate reduced by a factor of -2 pumps/min. In order to achieve the 14 pumps/min rate we would have to tweek the time.sleep() parameters a bit but consuming air at a reduced pressure would increase the flow rate significantly, triggering the compressor to turn on every 10 seconds. However, in case of a loose connection or leakage low pressure air can do less physical harm whereas high pressure air reduces noise pollution but optimizes the air flow for the same pump rates. We wish to test this aspect further to strike a good balance between safety and efficiency.

We confirmed with our Stakeholders that the compression rate achieved and the amount of torque provided by the pneumatic cylinder while squeezing the BVM ensured enough oxygen supply reached the patient’s lungs. We demonstrated the amount of air being pumped by attaching a balloon on the end of the expiratory valve of the BVM. The balloon inflated to a diameter of 9cm when I squeezed it by hand thoroughly and about 7cm when pumped using the automation. This experiment suggests that manual pumping can be irregular and unregulated which could lead to internal damage but a regulated system with controlled pumping is desirable as it is safer to use.

Automated Pumping

Manual Pumping

We tried designing a barometer out of spare plastic tubing to compare the difference in the force distribution between the piston head being mounted by a cushion versus the metal head being exposed. We were unable to complete this test as our homemade barometer did not end up working, the tubing available to us was extremely short so everytime the BVM would be squeezed even by hand, the water would spout out (even though we tried adding very tiny amounts).

The patient’s safety, convenience and comfort have been the backbone principles driving our innovation so when it came to implementing and operating the electrical components, we were equally cautious. We checked the voltage and power threshold for each component and made sure to measure the voltage and current passing through them, using a multimeter, before the circuits were made, preventing a short circuit.

Where would our project go next?

We designed the arms in a way that the automated mechanism could be brought to the patient in the ambulance instead of situating the patient close to a fixed bedside device. This would allow us to work around the medical infrastructure currently in place. For eg. if a patient requires intubation, the endotracheal tube inserted normally protrudes about 1 feet out of the patient’s mouth. In case a bedside ventilator is available on board, extra tubing would have to be connected to the endotracheal tube to enable secure connectivity between the device and the patient. While this might be a feasible solution for hospital setups, there are many health risks associated with the same for an ambulance journey. Improper fastening of the extended tube can lead to injuries especially when the ambulance is in motion. Thus, our design can bring much value in securing patient safety and comfort.

We were initially considering the installation of two arms, perpendicularly. One arm end would have a C-bracket clasp that would hold the BVM whereas the other arm end would have the pneumatic cylinder fastened. Upon consultation with our supervisor we comprehended that this design would not be able to withstand the structural load i.e it could generate shear and tension forces that would act upon the arm holding the BVM, as a consequence of force being applied in the perpendicular axis. After careful consideration we inferred that there should be a C-structure erected atop the arm such that the cylinder pumps in the direction of the arm axis.

Attached here is our video for the 3D design we would have implemented if not for the pandemic.

Technical Considerations:

Design Implementation

Attached is the 15-minute video for our design implementation.

Project Management

Although we are an agile team, for the winter semester we modified our project management style, using a waterfall-agile hybrid approach for the alpha release, and an agile approach for the beta and final release.

The alpha release consisted of exploring different design pumping mechanisms to automate the BVM. The sprints during the alpha release were mainly based on researching into the different pumping mechanisms and discuss the design ideas by consulting different industrial experts.

Here is a snapshot of the Alpha release Sprint Planning board:

The beta release had to be modified due to delays in the shipments of parts. The original beta release plan entailed completing the assembly of the device. But instead, we tried getting the design with whatever parts were available during that moment. The Rasberry Pi, and the RPi Relay board were available and so we decided to figure out how those two components work and how could we power them. Additionally 3D models were created for the arm design to mount the pneumatic cylinder and the BVM as well as the assembly of the pnuematic components encapsulated in a casing. We formulated sprints to complete these tasks, and allocated the tasks according to team members availability and skill sets.

Here is a snapshot of the Beta release roadmap Sprint Planning board:

The final release consisted of assembling and testing out the final design. Once we received the parts, we ran a sprint to allocate assignments for each task. We tested each part individually before connecting the whole design together such that it is easy to troubleshoot and figure out if any part is damaged in advance. We figured out how to power up the solenoid and the sensors during this release.

Here is a snapshot of the final release Sprint Planning board:

Work Distribution

Creating an even work distribution was one of the challenges of the capstone project. Although we had strategies that helped balance out the work somewhat ultimately at different points in the project the work distribution was skewed/non-uniform. We used an agile project management style. The assigned product owner created tasks on the asana and the scrum master assigned the task according to the team members’ access to resources in terms of access to the advisors,communication with stakeholders and physical access to the components of the product.

Since the assembly could only be done in one household and congregating to assemble the parts was not possible due to COVID-19 restrictions, the responsibility of assembling the parts would often fall on one group member. This happened for both the MVP and the final product assembly, and made it very challenging for group members to do work concurrently. To support the group members doing assembly, other group members would take on other responsibilities, such as buying and delivering various components for the product. We would also occasionally ferry the parts back and forth between households, according to the availability of each team member. If not for such crisis, it would be possible for the team to work together at one location. Unfortunately fate was not on our side in this regard.

Work distribution breakdown: (Not perfectly to scale but provide a rough estimate for the work distribution)

Financial Management

The final cost was $695.59 including the cost of the air compressor, which 108 ambulances are quite often equipped with. Without the air compressor this could be reduced to $520.46, which in our estimation is a more fair estimation of the cost of our product.

Our original design included multiple components that we did not include in the final design because we were afraid of going over the budget. The original price for all the components of the pumping mechanism was $1047.95, well over the $800 budget. The reason for this discrepancy was the difference between the final cost of the pneumatic components ($522.94) and the initial cost of the pneumatic components ($876.30). This amounted to a $352.36 difference, enough to bring the full cost under the budgetary requirement ($800 for the pumping component of the capstone project).

There were two main areas where we found cost savings. Firstly, the pressure sensor on the original air prep unit was removed. The Applications Engineer at Festo advised that although the pressure sensor could be useful for fault detection, it was not essential for the more crucial pumping functionality of the automatic BVM, and therefore not required for our design. Additionally, it was noted that it could easily be added on afterwards if the budget allowed for it. Secondly, the original Cylinder included a specially manufactured rectangular plate which included a cushion that would evenly distribute the force over the surface of the BVM causing smooth pumping and doing no damage to the self inflating bag. This cylinder was replaced by a round, more cost effective one. In order to emulate the cushioning effect that the original cylinder came with, we attached a sponge to the head of the piston, at no additional cost, and quite similar functionality. Together, these two items added up to $352.36. The lesson learned from this was that when developing a low cost mechanism, it is important to only include items that are absolutely essential, and to attempt to find substitutes wherever possible.

We saved money by either borrowing or recycling components as much as possible. In our final submission, all non-pneumatic components (with the exception of the relay board) were either borrowed or recycled, including the power supply, breadboard, Raspberry Pi. The cost of each of these devices is unknown, but it would be reasonable to assume some of these components like the power supply would be provided by the ambulance. The Raspberry Pi would increase the cost of the device by $47.45 if it was purchased brand new. However the Raspberry Pi is only needed for prototyping, as the programmed microcontroller is the only component needed for a mass produced design. This would not significantly increase the overall cost of each device.

We believe that the job we did saving money was quite admirable. Although our device cost about $ $695.59, this number could be reduced to $520.46 when one factors in that many of the 108 ambulances that our product was designed for include air compressors (ours cost $175.13.) This price could potentially be further reduced when one considers that when parts are bought in bulk for mass production, they are sold for less. Having our device be low cost was one of the key goals for our project and we believe we have met that at $520.46 we have met that requirement.

Link to team expenses document
Original cost of pneumatic components
Updated cost of pneumatic components

Lesson Learned/Reflection

Jonathan Blumenfled

Shivani Joshi

Priyam Shah

Areeba Abidi

Acknowledgment and Appreciation

We acknowledge, appreciate and thank everyone involved in helping our team succeed in our Capstone journey, you have all had a role in making this product possible.

Franz Newland: for taking us under his wings after a rushed proposal submission and approval at the start of them. You were a true mentor and support system for the team, we cannot appreciate all your technical, emotional and logistic support enough.

Stakeholders: for educating us on the circumstances under which the health care systems in the United States, Canada and India are operating, their distinct challenges and strengths. For always making time to consult on our medical and infrastructure related queries. For providing feedback on the practicality of our final design and so much more.

Nimesh Joshi: for explaining how pneumatics work, providing technical feedback on the feasibility and appropriateness of our final design in the mechanical engineering perspective. Thank you tons for helping us make an essential part of our 3D model, proof our concept once ready and always being a phone call away to provide all types of engineering support.

Florin Urban, Michael London and Eddy (Aztec Supply): for helping us actualize our design by proposing different approaches on implementation along with sourcing and delivering Festo products to our door at student discounted costs overnight. We can attribute the timely completion of our hardware implementations to your help.

Melina Tahami and Hannah Yorke Gambhir (Capstone team C): for helping us learn SolidWorks and get our final design 3D printed, thank you so much for taking out the time to help us out.

Course Directors, Claudia and TAs: We extremely appreciate all the considerations, feedback and support you provided us support throughout the term!

To all team member’s friends and family that loaned or gifted various household items and available electrical components to the team through our product iterations.

References

Icons made by wanicon from www.flaticon.com
Icons made by Freepik from www.flaticon.com
Icons made by Flat Icons from www.flaticon.com