Someone's life could be at risk if the IV Fluid Warmer does not operate properly. To better understand and address potential failure modes of the device, I have performed a system FMEA. Through this analysis I have developed a design strategy that includes double redundant fault detection and user alert methods.
Even so, there are a few failure modes where it is both difficult to detect the failure and difficult to convey to the user that a failure is occurring. These challenging failure modes include the scenario when either the microcontroller fails (a rare occurrence) or if the power being supplied to the device is interrupted (a likely occurrence particularly if a battery is used to power the device). Under these scenarios, if the user isn't actively watching the indicator lights on the device, they are unlikely to notice that the lights are no longer lit up and the device is no longer operating. There may be a 30 second or so grace period after either failure scenario when heat continues to transfer from the unit to the IV fluid, but eventually the rate of heat transfer to the fluid will slow and the IV fluid exiting the device may drop to an unsafe temperature. My current thought to alert the user during these scenarios is to either include a battery or capacitor in the device that will discharge to a buzzer when the microcontroller is no longer functioning (when it either fails or doesn't receive power).
Next I put together a list of system fault conditions that will implemented in software to detect when a failure mode has occurred. Hard faults will cut power to the heaters. Hard fault conditions represent scenarios where there is risk that the system will heat the IV fluid to 42 degC or greater and cause hemolysis of the blood products. A soft fault will just inform the user (in a sufficiently annoying manner) that the system isn't functioning properly and that it may or may not still be heating the IV fluid properly. The soft fault allows the user to recognize that there is an issue with the device and decide whether it is in the patient's best interest to continue to administer the fluid or to stop and investigate the cause of the fault. Three sensing devices will be used by the system for fault detection: four heater control temperature sensors, a system level hall current sensor, and an analog voltage sense.
ID
Description
Type
F01
erroneous temperature sensor reading
soft
F02
large variation of temperature sensor measurements
soft
F03
temperature sensor above max temperature limit
hard
F04
temperature sensor below min temperature limit
soft
F05
current is larger than max allowable current
hard
F06
increase in current is larger than max allowable current change
hard
F07
decrease in current is larger than max allowable current change
soft
F08
input supply voltage below min supply voltage
soft
F09
input supply voltage above max supply voltage
hard
Finally, I have decided on the following strategy for alerting the user of the state of the system using three colored LEDs and a buzzer.
I threw together a first iteration prototype of the enclosure today. It is comprised of a 250mm section of 3in diameter galvanized air duct, a male thread 3in PVC end cap, and a 3in compression pipe plug. The current plan is to mount the power switch and led indicator lights in the white PVC end cap. Eventually, I will create a permanent water tight seal between the white end cap and duct pipe. The red compression pipe plug is meant as a service port. It forms a water tight seal with the duct pipe when tightened and can easily be loosened in order to get access to the internal electronics for troubleshooting.
For the highest flow rate, the tube will need to be wrapped around the entire length of the duct pipe. Testing is needed to further understand how many wraps around the duct are required for the lower flows. Certainly wrapping around the entire length is most energy efficient, but may use more tube length than desired or result in a higher than acceptable pressure drop from inlet to outlet.
The silicone heater pads will wrap around the inside circumference of the metal duct pipe. Electronics will sit in the center core of the pipe (inside the wrapped silicone heater pads). I may route the 12V power supply connection through the duct pipe near the red compression pipe plug (with chord grips for a water tight seal).
The first iteration of the system architecture assumes two 100W silicone pad heaters inside a 3in diameter HVAC air duct. PID control on an Arduino Nano will be used to maintain constant temperature. Four thermistor temperature sensors will be used for temperature feedback. Heater power is controlled with two mosfet boards that are typically used for 3d printer heated beds. IV tubing is wrapped around the 3in air duct multiple times. Heat is generated by the silicone heater pads, transferred to the air duct, and then transferred through the IV tubing to the fluid.
The system will maintain temperature at 39 degC. If temperature exceeds 42 degC during normal operation, the heaters will turn off and the user will be informed through a visual and audible alarm. If the device temperature drops below 36 degC during normal operation, the user will be informed through a visual and audible alarm. Upon startup, the system will inform the user when the device has warmed up to 39 degC through a visual indicator. I may include input voltage sense for battery powered systems. When battery voltage drops below the voltage threshold a visual indicator would turn on to show low battery power. The voltage threshold is dependent on which battery system is used so would have to be configurable in the microcontroller code.
The latest cost breakdown shows a total system cost of $60 (excluding the required 12V, 20A power source). This cost estimate is not conservative. I expect the final design to be closer to $75.
A quick calculation shows that the maximum possible IV fluid flow rate with two 100W heaters (at 85% heat transfer efficiency) is approximately 80mL/min. This assumes that the blood product enters at 4degC (typical whole blood and packaged red blood cell storage temperature) and leaves at 39degC.
Assuming the electronics consume an additional 12W of power, I arrive at a duration between battery charges of approximately 1 hour for a typical 12V, 45Ah car battery at a 17.7A discharge rate (based on the lead acid battery voltage curves I found online). At 80 mL/min, it is theoretically possible to warm thirteen 350mL units of blood product before the battery needs to be recharged.
- blood warmers are rarely needed during routine transfusions, they are used when rapid transfusion of components is required - blood storage temperature, 2 to 6degC (whole blood/red cell component) - blood tranfusion temperature, 30 to 37degC - suggested rate of transfusion (adult) + whole blood 150-200mL/hr + PRBC (packed red blood cells) 100-150mL/hr + platelets/plasma 150-300mL/hr - suggested rate of transfusion (pediatric) + whole blood/PRBC, 2-5mL/kg/hr (17 mL/min allows transfusion in less than 30 minutes) + platelets/plasma, 1-2mL/minute - duration times for transfusion + whole blood/PRBC, less than 4 hrs (start within 30 minutes of removing from refrigerator) + platelet concentrate, less than 30 minutes (start immediately) - warmed blood is most commonly required for: + large volume rapid transfers * adults, more than 50mL/kg/hr * children, more than 15mL/kg/hr + exchange transfusions in infants + patients with clinically significant cold agglutinins (patient produces autoantibodies which cause agglutination of red cells at cold temperatures) + rapid infusion of blood products through central lines - blood should never be warmed in a bowl of hot water as this could lead to hemolysis (i.e. rupture) of the red cells which could be life-threatening when transfused - a radiant heater should never be used to warm the blood being transfused because of the risk of hemolysis - volume per unit of: + whole blood, 450ml + PRBC, 220-350ml + plasma, 180-270ml + platelets, 35ml - material properties of blood + density 1.05g/ml (1050 kg/m^3) + specific heat 3490 J/kg*K - medical tubing materials + polyvinyl chloride (PVC), most common (approx 30% of market) + polyethylene, second most common + thermoplastic elastomers (TPE) + nylon + silicone - blood transfusion IV tubing catheter size + adults, 18 gauge (nominal inside diameter 0.838 +/- 0.038mm) + pediatric, 22 gauge minimum (nominal inside diameter 0.413 +/- 0.019mm) - infusion/transfusion line diameter + outside diameter 4 to 7mm - warming blood to temperatures greater than 42degC may cause hemolysis - nominal capacity of a standard 12V car battery is approximately 45Ah
John Opsahl
I threw together a first iteration prototype of the enclosure today. It is comprised of a 250mm section of 3in diameter galvanized air duct, a male thread 3in PVC end cap, and a 3in compression pipe plug. The current plan is to mount the power switch and led indicator lights in the white PVC end cap. Eventually, I will create a permanent water tight seal between the white end cap and duct pipe. The red compression pipe plug is meant as a service port. It forms a water tight seal with the duct pipe when tightened and can easily be loosened in order to get access to the internal electronics for troubleshooting. 
