Optimum Pressurization Mechanism for a Non-Electrical Piston-Driven Infusion Pump

: Intravenous (IV) ﬂuid therapy is a common current medical practice, and the method has remained unchanged for more than a century. The IV bag is suspended from an IV stand or pole, and the pressure created by gravity is used to administer the drug. However, this method inevitably reduces the mobility of patients, and may cause accidents such as falls during movement. To solve these problems faced in home care, nursing home, and hospital settings, this study aims to develop a non-hanging, non-electrically driven IV infusion pump with reasonable portability and operability. In this study, instead of gravity, atmospheric pressure was used as the driving force. We developed several prototypes based on different pressurization mechanisms using vacuum piston cylinders as the driving mechanism, in order to ﬁnd an optimum mechanism capable of producing a stable ﬂow rate comparable to the suspended drip system. Tests on performance in terms of discharge ﬂow rate were conducted on three feasible prototypes, based on three different pressurization mechanisms, using a gravimetric test bench built for this purpose. The tests showed that the pressurization mechanism using an inﬂating sleeping-bag-shaped air bag to compress a drip bag achieved the best performance in terms of ﬂow rate stability.


Introduction
From treating acute stroke patients or patients with acute pain in emergency cases [1][2][3], to providing medical care for terminally ill and critical pediatric patients [4][5][6][7], to administering therapeutic treatment for home care patients [8,9], as well as managing nutritional support to enhance health recovery [10], intravenous (IV) fluid therapy is a common medical practice at all levels of urgency, with 80-90% of all patients receiving IV therapy during a hospital stay [11]. IV therapy allows the parenteral administration of a fluid substance (solution) directly into a patient's vein, mainly for three purposes: to provide fluids, to give drugs, or to transfuse blood products [11]. The first attempts at IV therapy can be dated back to the 17th century in which experiments were performed on animals using a device comprising a quill and a bladder [12]. It was not until the early 19th century, that practices on humans were first documented as curative treatments for hemorrhage during delivery [13] and for dehydration during the cholera epidemic [14,15]. While previous IV infusions were performed at a rapid pace, the slow and continuous drip method was first reported in the early 20th century [16]. IV therapy has since become established in modern healthcare and today is not limited to hospital settings, but is also implemented at home, in nursing centers and practitioners' clinics. The devices have evolved from a quill and bladder, to a disposable catheter and plastic bag. Advances in modern technologies such as microelectromechanical systems (MEMS) and the Internet of Things (IoT) have also prompted many studies applying sensing and wireless communication technologies to IV drip devices for the detection of IV fluid level [17][18][19], sensing of dosing flow rate [20][21][22] and remote monitoring of IV therapy [23][24][25][26]. Yet, the fundamental method of gravity drip infusion in which an IV bag is suspended (typically on an IV pole) at a constant height from a drip point (usually the antecubital vein of an elbow) to maintain an infusion pressure (derived from the gravitational potential) as shown in Figure 1a, has remained largely unchanged, mainly due to its unrivaled simplicity, practicality and cost-effectiveness.
While previous IV infusions were performed at a rapid pace, the slow and continuous drip method was first reported in the early 20th century [16]. IV therapy has since become established in modern healthcare and today is not limited to hospital settings, but is also implemented at home, in nursing centers and practitioners' clinics. The devices have evolved from a quill and bladder, to a disposable catheter and plastic bag. Advances in modern technologies such as microelectromechanical systems (MEMS) and the Internet of Things (IoT) have also prompted many studies applying sensing and wireless communication technologies to IV drip devices for the detection of IV fluid level [17][18][19], sensing of dosing flow rate [20][21][22] and remote monitoring of IV therapy [23][24][25][26]. Yet, the fundamental method of gravity drip infusion in which an IV bag is suspended (typically on an IV pole) at a constant height from a drip point (usually the antecubital vein of an elbow) to maintain an infusion pressure (derived from the gravitational potential) as shown in Figure 1a, has remained largely unchanged, mainly due to its unrivaled simplicity, practicality and cost-effectiveness. However, the gravity-driven method is imperfect: it confines a patient's movement, causing stress, especially for patients routinely undergoing long hours of slow drip such as IV chemotherapy for oncology patients [27]. For hospitalized athletes, limiting their mobility may also affect their physical and mental rehabilitation [28]. On the other hand, elderly patients who move with a wheeled IV pole, for instance, to attend toileting needs, may risk an accidental fall, resulting in unintentional injuries [29]. These problems of gravity drip have motivated much effort to produce inventive and innovative drip devices such as shoulder-mounted drip poles [30][31][32][33], a neck-mountable IV bag design such as a U-shaped pillow [34], wearables for carrying battery-powered infusion pumps [35][36][37], and even an autonomous IV stand that can monitor a patient's movement and move concurrently with the patient [38]. These studies provide solutions from different perspectives. Efforts in [30][31][32][33]38] focused on the IV pole, modifying it to become body-mountable [30][31][32] with automatic pole-balancing control [33], or sustaining the pole structure but causing it to move and follow the patient on its own [38]. The approach taken in [34,35- However, the gravity-driven method is imperfect: it confines a patient's movement, causing stress, especially for patients routinely undergoing long hours of slow drip such as IV chemotherapy for oncology patients [27]. For hospitalized athletes, limiting their mobility may also affect their physical and mental rehabilitation [28]. On the other hand, elderly patients who move with a wheeled IV pole, for instance, to attend toileting needs, may risk an accidental fall, resulting in unintentional injuries [29]. These problems of gravity drip have motivated much effort to produce inventive and innovative drip devices such as shoulder-mounted drip poles [30][31][32][33], a neck-mountable IV bag design such as a U-shaped pillow [34], wearables for carrying battery-powered infusion pumps [35][36][37], and even an autonomous IV stand that can monitor a patient's movement and move concurrently with the patient [38]. These studies provide solutions from different perspectives. Efforts in [30][31][32][33]38] focused on the IV pole, modifying it to become body-mountable [30][31][32] with automatic pole-balancing control [33], or sustaining the pole structure but causing it to move and follow the patient on its own [38]. The approach taken in [34][35][36][37] was to remove the IV pole and re-design the IV bag to make it attachable to the neck to gain gravity potential [34], or to maintain the current IV bag design but resort to battery-driven actuators for fluid delivery [35][36][37]. Echoing the motivation of these efforts, our aim was to find an alternative and sustainable solution by considering the following criteria: (i) trying to preserve the existing advantages of gravity drip which is easy to set up, electricity-free and highly affordable, (ii) placing priority on the stability of the infusion flow rate for treatment efficacy and patients' safety, and (iii) emphasizing compatibility with existing IV bags and infusion tube sets. With these criteria, we strove to achieve a concept of electricity-free and non-hanging IV drip therapy that will enhance resilience in medical treatments, especially at evacuation shelters in disaster-hit areas or refugee camps in war-stricken regions.
In this work, we ventured into a novel approach in which we removed the IV pole but sustained the current design of the IV bag, and at the same time, sought a non-electric way to drive the infusion fluid. Our aim was to find an alternative driving force to gravity, design and test feasible pressurization mechanisms acting on an IV bag to transform the driving force into fluid pressure for fluid infusion, and find the best-performing pressurization mechanism for the infusion pump. As an alternative to gravitational force, for instance, a driving force can be derived from the pressure difference between a vacuum and the atmospheric pressure in a vacuum piston cylinder [39][40][41][42][43][44]. Another feasible option could be an elastic force exerted by a spring [45,46] or a bellows [47]. For the pressurization mechanism, viable approaches can generally be categorized into three options, namely, (a) pressing a rigid plate on a drip bag [43,[45][46][47], (b) squeezing a drip bag by one or multiple rollers [48], and (c) compressing a drip bag by an inflating air bag [49][50][51][52]. However, it is unclear whether these mechanisms have been tested in terms of stability of infusion flow rate. First, we considered atmospheric pressure as a sustainable source of driving force on earth and, thus, an ideal alternative to gravity, leading us to employ the vacuum piston cylinder as the driving mechanism. Figure 1b presents a typical structure of a vacuum piston cylinder in which a piston separates a sealed vacuum chamber and an open atmospheric chamber. A driving force, derived from the pressure difference between the two chambers, acts towards the closed end of the cylinder. The magnitude and moving stroke of the resultant driving force are determined by the dimensions (i.e., diameter and length) of the cylinder. Second, we fabricated several novel and feasible prototypes combining one or multiple units of vacuum piston cylinder with each of the three pressurization mechanisms mentioned above. Third, to observe their potential in producing a stable infusion flow rate, we performed evaluation tests on the discharge flow rate from each prototype of infusion pump using a gravimetric test bench built for this purpose. Finally, a conclusion was drawn on the best pressurization mechanism, with prospects of future work.

Pressurization Mechanisms
We aimed to design an optimum pressurization mechanism acting on a drip bag, with the infusion pump producing a certain level of dosing stability and accuracy, in line with medical guidelines [53,54] and, ideally, comparable to the existing suspended drip system. In the succeeding sections, we discuss the problems and challenges faced in developing various pressurization mechanisms. At this early stage, design emphasis was placed more on prototype functionality rather than compactness and user-friendliness. Additionally, it should be noted that the outlet of the drip bag was orientated upwards in most of the prototypes for easy air-bleeding of the drip bag. Another noteworthy point is that all the infusion pump prototype designs were intended to be capable of generating a fluid pressure of approximately 15 kPa inside the drip bag, which is enough to overcome venous pressure.

Prototypes Using a Pressing Rigid Plate
A prototype (hereafter, V-PRESS) was fabricated, comprising a pressing plate orientated in a V-shaped layout against the housing wall (see Figure 2). The pressing mechanism was driven by a vacuum piston cylinder through a connecting structure. Theoretically, V-PRESS was designed to exert an increasing rotational momentum on the pressing plate as it moved towards the wall, with the aim of compensating the pressure (flow rate) drop during fluid infusion. However, measurements of the resulting fluid pressure from the drip bag did not result as expected, showing a significant pressure drop. The main reason for this was the change of contact surface area between the pressing plate and the drip bag as the mechanism moved. Placing an air bag between the pressing plate and the drip bag to improve the adherence of the contact surface did not solve the problem either. Figure 2c shows how to reset the pressing plate back to the original starting position.
drop during fluid infusion. However, measurements of the resulting fluid pressure from the drip bag did not result as expected, showing a significant pressure drop. The main reason for this was the change of contact surface area between the pressing plate and the drip bag as the mechanism moved. Placing an air bag between the pressing plate and the drip bag to improve the adherence of the contact surface did not solve the problem either. Figure 2c shows how to reset the pressing plate back to the original starting position. Meanwhile, we worked on another prototype, as shown in Figure 3. This prototype (hereafter, UP-PRESS) worked by elevating the pressing plate, using two vacuum piston cylinders, to exert pressure on the drip bag (see Figure 3a). This mechanism was thought to have the advantage of maintaining a more consistent contact area between the pressing plate and the drip bag. In Figure 3b, a stopper holds the pressing plate at the standby position before release. In Figure 3c, flipped-up handles allow the pressing plate to be pushed back to the standby position after operation.  Meanwhile, we worked on another prototype, as shown in Figure 3. This prototype (hereafter, UP-PRESS) worked by elevating the pressing plate, using two vacuum piston cylinders, to exert pressure on the drip bag (see Figure 3a). This mechanism was thought to have the advantage of maintaining a more consistent contact area between the pressing plate and the drip bag. In Figure 3b, a stopper holds the pressing plate at the standby position before release. In Figure 3c, flipped-up handles allow the pressing plate to be pushed back to the standby position after operation. reason for this was the change of contact surface area between the pressing plate and the drip bag as the mechanism moved. Placing an air bag between the pressing plate and the drip bag to improve the adherence of the contact surface did not solve the problem either. Figure 2c shows how to reset the pressing plate back to the original starting position. Meanwhile, we worked on another prototype, as shown in Figure 3. This prototype (hereafter, UP-PRESS) worked by elevating the pressing plate, using two vacuum piston cylinders, to exert pressure on the drip bag (see Figure 3a). This mechanism was thought to have the advantage of maintaining a more consistent contact area between the pressing plate and the drip bag. In Figure 3b, a stopper holds the pressing plate at the standby position before release. In Figure 3c, flipped-up handles allow the pressing plate to be pushed back to the standby position after operation.

Prototypes Using Squeezing Rollers
Another prototype (hereafter, MOVE-ROLLER) that worked by squeezing the drip bag using a pair of rollers, was also developed (see Figure 4). The rollers rotated themselves and squeezed a fixed-mounted drip bag while being pulled down by the vacuum piston cylinders (see Figure 4a,b). However, a problem arose in which the drip fluid escaped from the high-pressure side to the low-pressure side inside the drip bag, resulting in leakage through the small gap between the rollers (see Figure 4c,d).

Prototypes Using Squeezing Rollers
Another prototype (hereafter, MOVE-ROLLER) that worked by squeezing the drip bag using a pair of rollers, was also developed (see Figure 4). The rollers rotated themselves and squeezed a fixed-mounted drip bag while being pulled down by the vacuum piston cylinders (see Figure 4a,b). However, a problem arose in which the drip fluid escaped from the high-pressure side to the low-pressure side inside the drip bag, resulting in leakage through the small gap between the rollers (see Figure 4c,d). To solve the problem faced by the MOVE-ROLLER, we developed a modified prototype (hereafter, FIX-ROLLER) as shown in Figure 5, in which a big roller was employed for the purpose of winding up the emptied part of the drip bag, thus, preventing the drip fluid escape encountered in MOVE-ROLLER. The opposite end of the drip bag outlet was attached to the big roller before operation. As shown in Figure 5a, a wire connected to a vacuum piston cylinder pulls and rotates the big roller, driving the whole mechanism. One should take note that, for simple representation, the movement of the piston (horizontally to the right) in Figure 5a differs from the actual movement, which is upwards in Figure 5c. Compared with MOVE-ROLLER, the new features in FIX-ROLLER included all rollers rotating at fixed positions, and the outlet of the drip bag facing upwards. Exterior and interior views of FIX-ROLLER are shown in Figure 5b,c, respectively. The rolling mechanism is unlocked for release in Figure 5d, and then reset back to the standby position in Figure 5e. To solve the problem faced by the MOVE-ROLLER, we developed a modified prototype (hereafter, FIX-ROLLER) as shown in Figure 5, in which a big roller was employed for the purpose of winding up the emptied part of the drip bag, thus, preventing the drip fluid escape encountered in MOVE-ROLLER. The opposite end of the drip bag outlet was attached to the big roller before operation. As shown in Figure 5a, a wire connected to a vacuum piston cylinder pulls and rotates the big roller, driving the whole mechanism. One should take note that, for simple representation, the movement of the piston (horizontally to the right) in Figure 5a differs from the actual movement, which is upwards in Figure 5c. Compared with MOVE-ROLLER, the new features in FIX-ROLLER included all rollers rotating at fixed positions, and the outlet of the drip bag facing upwards. Exterior and interior views of FIX-ROLLER are shown in Figure 5b,c, respectively. The rolling mechanism is unlocked for release in Figure 5d, and then reset back to the standby position in Figure 5e.

Prototypes Using Inflated Air Bags
To identify which shape of air bag performed best as a pressurization mechanism, we conducted a series of tests (Tests No. 1-3) as shown in Figure 6. Test No. 1 acted as an observatory test without using any air bag (see Figure 6a,b). In all tests, two parallel plates used for compression were loaded with dead weight to exert a constant force. A drip bag and air bags were filled with air. The air was released from the drip bag at a low flow rate controlled by a valve, while the air in the air bag was trapped by closing the valve. Pressures in the drip bag and the air bag were measured simultaneously.
Test No. 1 showed the steady decreasing of the drip bag pressure that was normalized by the initial drip bag pressure, though under a constant force from the dead weight (see Figure 7a). This may have been due to the change of contact surface area between the plate and the drip bag as the drip bag deflated. The pressure value in the ordinate in Figure 7b,c was normalized by the initial air bag pressure. At the starting point in Test No. 2, the drip bag pressure was higher than the air bag pressure (see Figure 7b). This may be attributed to the surface tension occurring at both the curving (arc-like) side-surfaces of the drip bag which were not in contact with the air bags [55]. However, this did not happen in Test No. 3, in which the side-surfaces of the enwrapped drip bag were in contact with the sleeping-bag-shaped air bag (see Figure 7c). In both Tests No. 2 and No. 3, it can be observed that the air bag pressure decreased slightly due to the change of contact surface area with the plates as the air bags compressed the drip bag. Seeing that the pressures of both drip bag and air bag showed the best unison in Test No.3 (see Figure 7c), we concluded that the sleeping-bag-shaped air bag (see Figure 6e) was the best inflating mechanism to pressurize the drip bag. The rolling mechanism is unlocked by turning a toggle switch at the exterior side panel. (e) The rolling mechanism is reset to the standby position using a hexagonal wrench.

Prototypes Using Inflated Air Bags
To identify which shape of air bag performed best as a pressurization mechanism, we conducted a series of tests (Tests No. 1-3) as shown in Figure 6. Test No. 1 acted as an observatory test without using any air bag (see Figure 6a,b). In all tests, two parallel plates used for compression were loaded with dead weight to exert a constant force. A drip bag and air bags were filled with air. The air was released from the drip bag at a low flow rate controlled by a valve, while the air in the air bag was trapped by closing the valve. Pressures in the drip bag and the air bag were measured simultaneously.  The rolling mechanism is unlocked by turning a toggle switch at the exterior side panel. (e) The rolling mechanism is reset to the standby position using a hexagonal wrench.

Prototypes Using Inflated Air Bags
To identify which shape of air bag performed best as a pressurization mechanism, we conducted a series of tests (Tests No. 1-3) as shown in Figure 6. Test No. 1 acted as an observatory test without using any air bag (see Figure 6a,b). In all tests, two parallel plates used for compression were loaded with dead weight to exert a constant force. A drip bag and air bags were filled with air. The air was released from the drip bag at a low flow rate controlled by a valve, while the air in the air bag was trapped by closing the valve. Pressures in the drip bag and the air bag were measured simultaneously. Test No. 1 showed the steady decreasing of the drip bag pressure that was normalized by the initial drip bag pressure, though under a constant force from the dead weight (see Figure 7a). This may have been due to the change of contact surface area between the plate and the drip bag as the drip bag deflated. The pressure value in the ordinate in Figure 7b,c was normalized by the initial air bag pressure. At the starting point in Test No.2, the drip bag pressure was higher than the air bag pressure (see Figure 7b). This may be with the sleeping-bag-shaped air bag (see Figure 7c). In both Tests No.2 and No.3, it can be observed that the air bag pressure decreased slightly due to the change of contact surface area with the plates as the air bags compressed the drip bag. Seeing that the pressures of both drip bag and air bag showed the best unison in Test No.3 (see Figure 7c), we concluded that the sleeping-bag-shaped air bag (see Figure 6e) was the best inflating mechanism to pressurize the drip bag. A prototype (hereafter, AIR-BAG) using the sleeping-bag-shaped air bag as the pressurizing mechanism was then developed (see Figure 8). To inflate the air bag, we designed a novel air-compressor mechanism that was driven by a vacuum piston cylinder. As shown in Figure 8a Figure 8b shows the actual fabrication of the prototype with a transparent interior view. To push the vacuum and air pistons back to the starting position, a hand-operated inflator was used to pump air into the compressed air chamber of the air piston cylinder (see Figure 8c). A prototype (hereafter, AIR-BAG) using the sleeping-bag-shaped air bag as the pressurizing mechanism was then developed (see Figure 8). To inflate the air bag, we designed a novel air-compressor mechanism that was driven by a vacuum piston cylinder. As shown in Figure 8a, a small piston with a cross-sectional area of a S in a vacuum piston cylinder (upper cylinder) pushes a large piston with a cross-sectional area of a L through a connecting rod in an air piston cylinder (bottom cylinder), resulting in compressed air being supplied to an inflating air bag enwrapping a drip bag. The pressure of the compressed air, p c , (gauge pressure) can be approximately estimated as p c ≈ p atm (a S /a L ) where p atm is the atmospheric pressure. Figure 8b shows the actual fabrication of the prototype with a transparent interior view. To push the vacuum and air pistons back to the starting position, a hand-operated inflator was used to pump air into the compressed air chamber of the air piston cylinder (see Figure 8c).
face area with the plates as the air bags compressed the drip bag. Seeing that the pressures of both drip bag and air bag showed the best unison in Test No.3 (see Figure 7c), we concluded that the sleeping-bag-shaped air bag (see Figure 6e) was the best inflating mechanism to pressurize the drip bag. A prototype (hereafter, AIR-BAG) using the sleeping-bag-shaped air bag as the pressurizing mechanism was then developed (see Figure 8). To inflate the air bag, we designed a novel air-compressor mechanism that was driven by a vacuum piston cylinder. As shown in Figure 8a Figure 8b shows the actual fabrication of the prototype with a transparent interior view. To push the vacuum and air pistons back to the starting position, a hand-operated inflator was used to pump air into the compressed air chamber of the air piston cylinder (see Figure 8c).

Evaluation of Flow Performance
As one of the criteria of being a possible alternative to the gravity-driven infusion method, the atmospheric pressure-driven method needs to achieve a dosing stability level comparable to the gravity method. Medical guidelines [53,54] also recommend that the flow rate accuracy for non-electrically driven infusion devices be within ±20% of the nominal flow rate. To conduct measurements of the discharge flow rate from the prototypes, we built a gravimetric test bench, which is discussed in the following section.

Gravimetric Test Bench
As shown in Figure 9a, a gravimetric (weighing) test bench was enclosed by a wind shield to avoid any wind effect. A pneumatically liftable stage supported a weighing balance at a variable height for flow rate measurement at different head pressures. Figure 9b shows a specially designed weighing vessel sitting on a weighing plate. The weighing range of the weighing balance was up to 6 kg, with a resolution of 0.01 g. As shown in Figure 9c, the weighing vessel had a special feature, in which an inner vessel stood on the bottom surface of the weighing vessel. One method of avoiding liquid dropping from the tube into the weighing vessel could be to immerse the tube into the liquid, but this would create immersion effects, such as a buoyancy effect, of the immersed tube beneath the rising liquid level. To maintain a constant immersed depth of the tube, many studies have adopted an overflowing inner vessel where a tube or a needle is immersed into the overflowing inner vessel [56][57][58][59]. Another method could be to maintain a constant liquid bridge or stream connecting the tube outlet and a liquid-absorbing material, such as glass filter or foam mounted inside a weighing vessel [60]. For liquid collection in this study, we combined the two methods, by not overflowing the inner vessel but filling it with liquid-absorbing spongy material so that the liquid stream from the tube outlet was absorbed. Liquid then wove through the sponge, down the inner vessel, and flowed out from the openings at the bottom of the inner vessel (see Figure 9c).

Tests of Flow Performance
Using the test bench described in Section 3.1, tests of flow performance were conducted on three feasible prototypes, namely: a. UP-PRESS, b. FIX-ROLLER and c. AIR-BAG, of which the measurement results are shown in Figure 10a-c. The left ordinate corresponds to the discharge flow rate, whereas the right ordinate corresponds to the amount of discharged liquid. In the performance test, liquid was discharged from drip bags with 500 mL content over a time period of about 2 to 3 h, at a flow rate adjusted between 3 and 4 g/min. Apparently, prototype AIR-BAG (see Figure 10c) showed a relatively high flow stability, compared with prototypes UP-PRESS and FIX-ROLLER (see Figure 10a,b). In terms of flow variation, UP-PRESS showed a relatively large up-and-down, while FIX-ROLLER demonstrated small zigzagging of flow fluctuation. In a general sense, the inconsistent contact surface interaction between the pressurization mechanism and the drip bag in the two prototypes accounted for these flow behaviors. The dynamic weighing method was adopted for the test bench, in which the increasing liquid mass collected inside the weighing vessel was measured by the weighing balance as a function of time, to ultimately obtain the mass flow rate. The flow rate was determined by a simple ratio of the differences in mass and time over a time interval. Though fitting methods such as linear fit over a time window will give a smoother curve with reduced noise, we considered that the noise was not so significant in the gravimetric system, and the present calculation method was good enough to see the flow performance of the infusion pump prototypes. Taking into consideration uncertainty factors such as liquid evaporation, resolution of weighing balance, buoyancy effect and so on, we performed a preliminary uncertainty analysis of the test bench, obtaining an expanded uncertainty of about 3.3% for the smallest flow rate of 3.0 g/min. The dominant uncertainty factor in the flow rate determination was the resolution of the weighing balance. Nevertheless, the amplitude order of the measurement uncertainty (3.3%) was considered sufficient in the context of the present study, seeing that the flow rate accuracy requirement for non-electrically driven infusion pumps is ±20%.

Tests of Flow Performance
Using the test bench described in Section 3.1, tests of flow performance were conducted on three feasible prototypes, namely: a. UP-PRESS, b. FIX-ROLLER and c. AIR-BAG, of which the measurement results are shown in Figure 10a-c. The left ordinate corresponds to the discharge flow rate, whereas the right ordinate corresponds to the amount of discharged liquid. In the performance test, liquid was discharged from drip bags with 500 mL content over a time period of about 2 to 3 h, at a flow rate adjusted between 3 and 4 g/min. Apparently, prototype AIR-BAG (see Figure 10c) showed a relatively high flow stability, compared with prototypes UP-PRESS and FIX-ROLLER (see Figure 10a,b). In terms of flow variation, UP-PRESS showed a relatively large up-and-down, while FIX-ROLLER demonstrated small zigzagging of flow fluctuation. In a general sense, the inconsistent contact surface interaction between the pressurization mechanism and the drip bag in the two prototypes accounted for these flow behaviors.

Discussion
In the previous section, we observed that the AIR-BAG prototype showed the most promising performance in terms of flow rate stability. To see how AIR-BAG compared

Discussion
In the previous section, we observed that the AIR-BAG prototype showed the most promising performance in terms of flow rate stability. To see how AIR-BAG compared with the existing gravity-driven method (pole-suspended drip system), we plotted the flow measurement results obtained for both methods in Figure 11. The left and right ordinates correspond to the discharge flow rate and the amount of discharged liquid, respectively, whereas the abscissa shows the measurement time. All quantities were represented in normalized and non-dimensional values to ease comparison. To avoid curves' overlapping that may cause confusion, only part of the curves (in orange and yellow) for the amounts of discharged liquid are shown in Figure 11. Appl. Sci. 2022, 12, 8421 11 of 14 was discharged from the drip bag, the flow rates infused by the AIR-BAG and the gravity decreased to 0.68 (68%) and 0.93 (93%), respectively, of the initial flow rates. In other words, during the majority (up to 90%) of the infusion process, the gravity drip method maintained a consistent level, in terms of flow rate, that was a maximum of 25% better than the AIR-BAG prototype. This prompted us to launch investigatory work into the next phase of study, to find ways to improve the AIR-BAG. Other than flow performance, much work has still to be conducted on developing the portability and ergonomics of the design. It should also be noted that the performance index described above had a scatter of approximately 10% in terms of the difference between the maximum and minimum values from the repeated measurements. Figure 11. Comparison of the infusion flow rate stability between the prototype AIR-BAG and the existing gravity-driven method.
Of note, the flow rates were adjusted by using the roller clamp and by monitoring the liquid drops in the drip chamber of the infusion tube set (see Figure 1a), which made it difficult to accurately adjust the flow rate. This resulted in certain discrepancies in flow rates between measurement runs. It may appear incomprehensible to observe that a quickly decreasing flow rate from the AIR-BAG (blue curve in Figure 11) produced a higher amount of discharged liquid (orange curve in Figure 11) than the highly consistent flow rate by the gravity method. This occurred because the actual (non-normalized) flow rate of AIR-BAG was slightly higher than that of the gravity method, resulting in a faster accumulation of discharged liquid in the weighing vessel.
The limitation of the infusion pump mechanism proposed in this study in general, as well as the limitation of the flow performance results discussed in the present context specifically, should also be noted. Since atmospheric pressure is the source of the driving force for the pressurization mechanism, the device performance is expected to vary with different altitudes; the difference may be especially significant between sea level and highlands or in an air ambulance or airplane. It is possible that thinner air will lower the maximum infusion flow rate of the infusion pump. This may require the need to devise 'infusion pump models for high-altitude use' in the future. Overall, as demonstrated in Figure 11, the discharge flow rate of the AIR-BAG decreased faster over the time span of measurement, than the gravity method. As a comparison index, we chose the value 0.9 (or 90%) in terms of the amount of discharged liquid in the right ordinate and traced it to the corresponding discharge flow rates in the left ordinate (see dotted and dashed grey lines in Figure 11). When 90% of the amount of liquid was discharged from the drip bag, the flow rates infused by the AIR-BAG and the gravity decreased to 0.68 (68%) and 0.93 (93%), respectively, of the initial flow rates. In other words, during the majority (up to 90%) of the infusion process, the gravity drip method maintained a consistent level, in terms of flow rate, that was a maximum of 25% better than the AIR-BAG prototype. This prompted us to launch investigatory work into the next phase of study, to find ways to improve the AIR-BAG. Other than flow performance, much work has still to be conducted on developing the portability and ergonomics of the design. It should also be noted that the performance index described above had a scatter of approximately 10% in terms of the difference between the maximum and minimum values from the repeated measurements.
Of note, the flow rates were adjusted by using the roller clamp and by monitoring the liquid drops in the drip chamber of the infusion tube set (see Figure 1a), which made it difficult to accurately adjust the flow rate. This resulted in certain discrepancies in flow rates between measurement runs. It may appear incomprehensible to observe that a quickly decreasing flow rate from the AIR-BAG (blue curve in Figure 11) produced a higher amount of discharged liquid (orange curve in Figure 11) than the highly consistent flow rate by the gravity method. This occurred because the actual (non-normalized) flow rate of AIR-BAG was slightly higher than that of the gravity method, resulting in a faster accumulation of discharged liquid in the weighing vessel.
The limitation of the infusion pump mechanism proposed in this study in general, as well as the limitation of the flow performance results discussed in the present context specifically, should also be noted. Since atmospheric pressure is the source of the driving force for the pressurization mechanism, the device performance is expected to vary with different altitudes; the difference may be especially significant between sea level and highlands or in an air ambulance or airplane. It is possible that thinner air will lower the maximum infusion flow rate of the infusion pump. This may require the need to devise 'infusion pump models for high-altitude use' in the future.

Summary and Conclusions
As a potential alternative to the gravity-driven infusion method, several prototypes based on three different pressurization mechanisms driven by a vacuum piston cylinder were designed and fabricated. The performance of three feasible prototypes, in terms of flow rate stability, were tested using a gravimetric test bench. Prototype AIR-BAG showed the most promising result in terms of flow stability compared with prototypes UP-PRESS and FIX-ROLLER. In the future, we will continue working on AIR-BAG to enhance its performance to a level comparable with the gravity-driven method.

Patents
Some of the infusion pump prototypes described in this paper are patent pending.