Ultra-Tall Synthetic Tree
For my final year at Virginia Tech, I worked on the Ultra Tall Synthetic Tree team. Our goal was to build the world’s tallest artificial tree, utilizing artificial transpiration. Transpiration is the mechanism trees use to transport water from their roots to leaves and will allow us to passively pump water and possibly use that potential energy to harvest clean energy in the future. From a physics standpoint, there is a nano-porous plate in the upper “leaf” portion of which utilizes Kelvin and Laplace pressures in combination with evaporation to create a negative pressure at the top and draw water from the bottom reservoir. My contributions to this project reside in the solar concentrator, and the artificial leaf. The most rewarding part of this project has been leading my team members and playing to everyone’s strengths. I served as the team facilitator, and was able to help and guide everyone to completing the design of this project with many novel, innovative changes. For this project, at the very beginning I categorized everyone into being either practical or theoretical. This allowed me to make sure there was at least one person of each “personality type” on each sub-team. I largely credit the cohesive nature of the team to this in the sense that the designs are very innovative, yet practical because everyone has someone with a complementary perspective to help them design each part. We need the theoretical people to help make sure what we are doing will function as intended, and we need practical people to make sure the design is manufacturable and can be assembled properly. Furthermore, little things such as weekly emails, sub team representatives, and honest opinions have carried this team far. I've put some content from our summary report, for a closer look at what we did.
Introduction
The goal of this project was to create the world’s tallest synthetic tree ever to demonstrate the scalability of synthetic trees and investigate their potential for pumped storage hydropower and other similar applications. Natural trees require a large amount of water to be pumped from the ground, up their trunks, and into their leaves for transpiration. A common belief is that the pumping of water is due to capillary rise, however this effect is not enough to explain the high flow rates present in most trees. Instead, the movement of water is due to a negative pressure generated naturally in the pores of each leaf. The evaporation of water causes a vacuum pressure, which propagates down the length of the tree due to the intermolecular forces present in water. This negative pressure is powerful enough to transport water up hundreds of feet in certain trees, such as the California redwood .
A project team previously replicated the Laplace pressure found in natural trees in a 10-foot tall synthetic device. A nanoporous ceramic disc acted as the nanoporous leaf, generating the negative pressure, which pulled water up from a lower reservoir through plastic tubes. Now that the concept has been proven, our team’s goal was to demonstrate that this technology is scalable by creating a tree that is above 50 feet in height. Being a scaled version of the previous project, our team had certain design challenges not faced by last year’s team. To begin the pumping process, each tube of the tree must be fully saturated with degassed water so the negative pressure can permeate the full length of the system. This was simple for the 10-foot version, as submerging the tubes in a pot of boiled water sufficed for the saturation process. Since this year’s version rendered this solution impractical due to size constraints, our team had to find a novel method to fill the tubes and nanoporous disc with degassed water.
While the leaf alone should provide a large ceiling for water flow rate based on Laplace pressure, the limiting factor of this flow rate will be the rate of evaporation (see “System Analysis and Testing”). Considering the larger scale, this year’s version introduced a solar concentrator to increase the rate of transpiration within our leaf. Demonstrating the success of this solar concentrator would help to show that synthetic trees are in fact scalable; larger versions of the device will simply require more (or bigger) solar concentrators. Therefore, solving the challenges that arise due to scale, both foreseen and unforeseen, was the primary purpose of this year’s synthetic tree project. As a team, we engineered a 50-foot tall synthetic tree, which was successfully charged via a pressure charging system and included a solar collector to increase evaporation rate.
Figure 1: Concept of previous team's project with the transpiration process shown on the right
Concept Generation, Analysis, Prototyping, and Concept Down Selection
The team determined that dividing the system into subcomponents would maximize efficiency and allow the team to complete research in a more timely manner. We separated the system into three main systems: tree charging system, leaf, and solar concentrator. After collecting information from our advisor regarding the function of each system, as well as conducting personal research within the respective subsystems, we were able to generate concepts that met both our customer needs and target specifications. We used multiple methods demonstrated in the lectures this semester, including the 3-5-7 method and mind mapping. We also created a morphological chart, which is shown in Figure 2, to determine the design solutions of the critical functions of our system.
Figure 2: Morphological Chart
Mind Map
A. Tree Charging System Concept Generation
It is necessary in our project to have the tree initially filled with water before starting the transpiration process. For the smaller scaled tree designed in the previous year, the tree was filled with water by using a small container of boiled water at the bottom of the tree. Because of the larger scale of our prototype, water was pumped from the bottom of the tree. We initially barred the idea of using an air pump, as the price of the pump would be relatively high, and the electrical energy required to pump the water up might be inaccessible from the location of the tree. Therefore, another system was developed; this “pressure charging system” consists of a large reservoir containing degassed water, connected from one side to a smaller water vessel and the bottom of the tree, and a nitrogen tank on the other side. A release valve would then be opened so the nitrogen pressurizes the large water vessel, and thus pumps the water up the tubes of the tree. The water will be forced from the large container to the smaller reservoir, before being pushed up the tubes until the hydrostatic pressure matches the pressure in the large tank. The initial pressure charging system imagined is represented by the drawing in Figure 3.
Figure 3: Early Drawing of the Pressure Charging System
However, during the building phase, the team decided to opt for an air compressor instead of a nitrogen tank. This will be shown and described later in the report. It is crucial for the success of our project, as mentioned earlier, that our system stays completely airtight and does not allow any air bubbles inside. The incorporation of compression fittings, which are also compatible with plastic tubing, in the adapter design was intended to ensure airtightness.
B. Leaf Concept Generation
To generate concepts for the leaf, we broke the system down into smaller functional groups, dictated by the leaf’s necessary functions, and generated different concepts for each group. These concepts build on but differ from the work done by the previous design team due to the much larger scale of our project. The leaf’s primary purpose, to generate negative Laplace pressure, is accomplished in a nanoporous disc at the top of the leaf. To allow for a continuous channel of water from the nanoporous layer through to tubes coming up from our ground reservoir, we generated multiple concepts for a “scaffold,” which also needed to integrate a channel to receive heat transfer from our solar concentrator in order to increase our transpiration rate. Based on the experiences of the previous design team, our faculty advisor recommended we used compression fittings as opposed to barbed fittings with sealant or another solution for our tubing, so the scaffold also incorporated nubs to which these fittings could be attached. Finally, the team generated different concepts to join the two layers of the leaf together. This preliminary design of the leaf system is illustrated in Figure 4 below. However, as the team continued into the spring semester, this design incorporating a heated channel was discarded due to the alteration of the solar concentrator design.
Figure 4: Leaf components used for concept generation
For the nanoporous disc, the team investigated multiple options: the purchase of a manufactured ceramic disc, manufacturing of a disc by 3D-printing or sintering, or direct deposition of a nanoporous film on our scaffold using a Chemical Vapor Deposition (CVD) machine. After communicating with a representative from the Oak Ridge National Laboratory, where we have potential access to a CVD machine, we learned that this process could only practically be applied to a material with holes or pores on a microscopic scale, which might limit the concept’s compatibility with different scaffold options. This contact did recommend that we deposit a layer of nanopores on a commercially available metal microporous mesh to get around that complication, so we added a new concept based on this design approach.
For the scaffold connecting the nanoporous disc to the tubes, our advisor originally suggested a microporous ceramic or metal cylinder into which we would machine nubs and a heating channel, sealed to the nanoporous disc with simple silicone sealant. To these concepts we added a solid metal scaffold into which we would drill channels directly from the tube nubs to the nanoporous disc. This would reduce the complexity of the system by removing the need for sealing the outer edge of the scaffold (potentially reducing the risk of failure of the seal), and it should allow for greater heat transfer to the transpiration surface as compared to a purchased ceramic microporous material. It would also be less brittle than a purchased ceramic microporous material. However, because we were concerned about uneven filling of the nanoporous disc during the charging process, we also considered a similar concept that included an interconnection between the parallel channels near the nanoporous layer. Depending on which combination of nanoporous disc and scaffold material we choose, silicone sealant alone could be used to join the two layers. Our other concept for joining the layers, a clamp combined with rubber O-rings and sealant, could provide increased security with regard to the airtightness of the system, but might be incompatible with more brittle materials required for some nanopore or scaffold concepts.
C. Solar Concentrator Concept Generation
In natural trees, there are hundreds, if not thousands of leaves that pull water up the tree compared to our design which only has one. Given that all of the water within our system was concentrated in this leaf, introducing a method for increasing evaporation from the nanoporous medium would be highly beneficial and could be done by using a solar concentrator. A solar concentrator is a device that reflects or absorbs sunlight into a designated surface area to be used for heating purposes. For our system, it was proposed that conductive fluid be pumped through the microporous scaffold of the lead. The fluid would be heated by the sun and focused on a tube, either plastic or metal, that is fixed in the scaffold to the leaf. It was crucial to use a solar concentrator in this year’s system, as we intended to show the scalability of flow rates within the tubes with a higher transpiration rate.
We determined that two methods of solar concentration would be examined: reflection and absorption. Utilizing reflective mirrors to concentrate heat onto a specific surface area could increase the total heat absorbed, however it would be quite costly and difficult to construct. Considering this, we produced a new concept that would use a black body to absorb sunlight and radiate heat over a surface. To transfer the heat, we found that using a high temperature-rated tube would be the most straightforward and effective solution, as we have to heat the interior of our microporous scaffold. In the reflective mirror design, the tube would be fixed at a calculated distance from the reflective surface, while for the black body design, the tube would be fixed inside and allowed to heat up. To transport the heat to the tube, we evaluated multiple materials and found that a glycol solution would be appropriate due to its high thermal conductivity and resistance to boiling. Car engine coolant uses a 50/50 mixture of glycol and water and has a high temperature rating, so we determined that it would be the best commercially available option. We also concluded that a low maintenance, high temperature chemical pump would need to be purchased to circulate the fluid. As will be discussed in the “Detailed Design” section, the team chose to forego the incorporation of a heating element within the leaf scaffold as we found this would complicate our design and was not practical for our purposes.
A parabolic concentrator as well as a trough reflector were evaluated as potential design solutions. The two concepts function quite similarly; the sun is reflected off of the mirrors and concentrated into a horizontal running tube that carries a thermal fluid. These designs are highly efficient as they produce large amounts of energy, yet they are both costly and difficult to maintain. Because the parabolic and trough reflectors need to always be directly facing the sun to achieve favorable efficiency, we considered a parabolic concentrator that tracks the sun throughout the day. It would utilize either a timer or optical device that notes the time of day and accordingly adjusts the angle of the device. A conical, or vertical cylinder design was also proposed; this device would work well at peak day, but again, the geometry limits the efficiency of the device significantly. Taking these different findings into account, we developed a new design that we call the planar solar concentrator. It was a black box that contained a coiled tube through which a pump would drive the thermal fluid. This system does not require tracking of the sun and is not limited in its shape. It is also very cost effective as well as simple to use, and demonstrates similar thermal results to the others due to its excellent average energy production. It was also proposed that adding reflective mirrors on the outside at both the east and west positions of the box would increase the amount of energy that could be absorbed by the system. To quantify the concepts and determine which would be best for further improvements, we performed concept down selection by weighting engineering characteristics using pairwise comparison and finding the concept with the best overall score.
Figure 5: Solar Concentrator Concepts
System Analysis and Testing
A. Governing Equations
The tree is governed by an equilibrium of pressures; while the negative Laplace pressure pulls the water into the leaf, it is being counteracted by the viscous and hydrostatic forces of the water. Therefore, the pressure equilibrium is described as:
PG is the hydrostatic pressure, PL is Laplace pressure, and PD is the Darcy pressure. The hydrostatic pressure, which is essentially the pressure due to the weight of the water, is described by:
where ρ is the density of water, g is gravitational acceleration, and H is tube height. This value was calculated to be 149 kPa using a tube height of roughly 15.2 meters.
To determine the pressure drop across the leaf, Darcy’s equation below was used.
Q is the volumetric flow rate of water through the leaf, t is the thickness of the nanoporous layer, κ is the intrinsic permeability, and A is the leaf area.
φ is the porosity of the nanopores, μ is the dynamic viscosity, and 𝜏 is the tortuosity. The porosity value was assumed to be 0.30 based on research of nano-scale pore occupation on a surface and the tortuosity was assumed to be 0.32 based on research into light permeation into silica membrane. Similar to the Laplace pressure, the intrinsic permeability is a function of the pore radius, while also being a function of the nanopore thickness. The nanoporous silica oxide will be deposited onto a commercially available microporous mesh. Because the layer is extremely thin, we have taken the thickness of the porous layer to be the thickness of the mesh, 0.0635mm.
The viscous pressure drop through the tubes was found using Poiseuille's equation below.
H is the height of the tubes, N is the number of tubes, and R is the tube radius. With H being set at 17 m, N being 7 tubes, and R being ¼ inch, the pressure drop from the Poiseuille equation is 3.39 * 109 * Q.
The Laplace pressure, which balances the three pressure drops mentioned previously, is generated by the leaf according to the equation:
γ is the surface tension of water, Ө is the contact angle of each menisci, and rpore is the pore radius. The Laplace pressure is required to maintain the menisci at the nanoporous layer. The surface tension was taken as 7.28*10-2 N/m; the major design factor that our group can control is the pore radius.
The Laplace pressure varies linearly with the transpiration rate, assuming the pressure is still below the maximum Laplace threshold (Ө = 0). While the negative pressure always has to counteract the hydrostatic pressure, the Pousielle and Darcy pressures are proportional to the flow rate of water in the tree. Because of conservation of mass, the transpiration rate equals the mass flow rate of water in the tubes. Figure 11 shows the relationship between transpiration and Laplace pressure for a tree of our design parameters. One can see that the Laplace pressure being generated during a particularly sunny day, -1.4 MPa, is much less than the maximum possible negative pressure of -14.3 MPa. This means that a similar leaf design (r = 10 nm) can even support a taller tree with faster flow rates than what we tested. This is a positive sign for potential scalability of the system.
Figure 11: Laplace pressure generation in response to transpiration rate
B. Solar Concentrator Contributions
The goal of the solar concentrator is to increase the transpiration rate of the system by increasing the local temperature and evaporating the water on the surface of the leaf faster. To relate the volumetric flow rate and the performance of the solar concentrator, the following equations were utilized.
Mdot is the mass flow rate of water, Wsolar is the power generated by the solar concentrator, and l is the latent heat of evaporation. Wsun is the solar power exerted on the surface of the Earth, on average, Ar is the reflector area, Ө is the reflector angle, and As is the surface area of the box surface. Based on our current solar box design, we expect to generate and direct 200 W of heat onto the leaf scaffold during daytime hours. This is based on the solar power concentrator directly into the box from the sun and from each of the reflectors. Assuming most of this heat is absorbed by the water, we would expect a mass flow rate of about 2.7 * 10-5 kg/s. Once again, the viscous pressure generated by this flow rate is lower than the maximum Laplace pressure, so it can be supported by the design.
Detailed System Design
With further research, ideation, and conceptualization, each subcomponent was solidified into one final concept. After utilizing the concept down selection, each subgroup was able to determine the ideal combination of features, and then further refine that to fit the compatibility needs of the other parts in order to make sure that the overall system worked adequately. The team additionally engineered a data collection interface and a mounting system for the device on the garage.
A. Pressure Vessel
The pressure vessel was mostly designed by students working under Dr. Boreyko in the 2019 - 2020 school year. This team made a few changes, but kept the same functionality. As stated earlier, the team did not choose to use a nitrogen gas tank, and instead invested in a commercial air compressor; not only is this cheaper, it allows for a greater customization and regulation of pressure into the tree and does not require any certifications. The air compressor pressurizes the large pressure vessel, and pushes water all the way up the system to the leaf. The team’s choice of a commercially available pressure vessel can be seen in Figure 12 below. This provided the advantage of a decreased build time and ready-made fittings to attach the tube going to the smaller machined pressure vessel. Furthermore, there is a pressure release valve to keep the pressure in the tank below 100 psi. A full CAD model can be seen in Figure 12 below.
Figure 12: Pressure Vessel CAD Model
B. Leaf
The concept that emerged as a preliminary front-runner through the concept downselection process is shown in Figure 13 below. This concept uses the machined/drilled metal scaffold with through-channels for the tubes to supply water to the nanoporous plate, with a common diffusion area above the channels to help the water enter the nanoporous plate evenly with no pockets of trapped air. This basic concept, a machined scaffold including a bolted clamp, is the only design that we can be confident will withstand the pressure of the clamp and of the negative pressure water without risk of damage, and also allows the largest possible disc area (and therefore tube width and number of tubes) of the options we found. Its only drawback relative to other designs, based on our preliminary analysis, is potentially the cost, but cost is a lower priority than some of the other engineering characteristics the leaf will effect, and the cost still should not exceed the other concepts by too much.
Figure 13: Preliminary Leaf Concept Schematic
However, upon additional discussion and feedback, a different design was proposed. This was the design that we decided to use moving forward; it used the same basic principle of clamping the nanoporous leaf to the bottom machined scaffold, however this time the clamp was bolted into the machined scaffold. In addition, the array of water tubing was changed to accommodate the possibility that water would not reach some areas of the nanoporous plate. Shown in Figures 14, 15 and 16 are the CAD of this final design. Figure 14 shows the leaf scaffold, which was designed to connect to the tubes via compression fittings. Figure 15 and 16 depict the components used to secure the mesh to the surface of the scaffold. The microporous mesh was cut to size and epoxied to the inner ring of the disc in Figure 16. This was then secured in the mesh clamp shown in Figure 15 and bolted onto the scaffold. We utilized a three part system due to the scale of the overall system. The clamp would permit airtightness of the leaf, which was our number one target specification, and the scaffold would serve as a channel for the water to travel from the tubes to the surface of the leaf.
Figure 14: Final Leaf Scaffold CAD
Figure 15: Final Leaf Clamp CAD
Figure 16: Final Leaf Mesh Disc CAD
C. Solar Concentrator
The final solar concentrator design was very similar to the initial design, but combined many elements of initial concepts into the final design. After finalizing the concept down selection, we determined that the utilization of an intermediate source of heat transfer to the leaf via tubes would be highly inefficient. Inspiration was taken from both the parabolic mirror and planar concentrator designs and was combined into a single design. The finalized design no longer used tubing and instead integrated the leaf directly into the concentrator. The concentrator box was made to be 2 square feet and the interior was painted black. Four reflectors were secured around the box at a 15 degree angle to reflect light onto the surface of the leaf and into the box. A thin plexiglass sheet was set on top of the box with a hole cut in the middle so the top of the leaf was exposed to the atmosphere. This allowed heat to be retained while permitting water to freely evaporate into the atmosphere. The solar concentrator was fixed onto an elevated table surface so the tubes leading into the bottom of the leaf could be properly routed over the top of the parking garage and into the solar concentrator. The final design is shown in Figures 17 and 18.
Figure 17: Planar Solar Module with Reflectors
Figure 18: Planar Solar Module with Reflector CAD
The main advantages of this concept were that heat was introduced into the leaf in a direct fashion, as opposed to running tubes through the concentrator and then circulating the fluid in the tubes from the concentrator to the leaf. For this particular design, the size of the concentrator was smaller, making it optimal for mounting the leaf inside. This also streamlined the design as the leaf and solar concentrator mounting structures were combined into one. The only drawback to this design was that it was less effective as the design scaled. If more heat was needed it would have been more ideal to revert to the design with the tubes running through the box. On the other hand, the solar concentrator wasn’t exactly a key component of the long term design of the tree, but rather a mechanism for its function. It was present to provide heat, meaning that it could be replaced by waste heat from a power generator.
Testing and Validation
During the semester, we developed a testing plan to verify the engineering characteristics of the components of our system. These target specifications are present in Table 1 below, associated with the instruments that were used to measure them.
We utilized the testing procedure above after the system was constructed. First, we tested the airtightness of the system by pushing air and water using the air compressor. We discovered leaks at the compression fittings and tubes, so we disconnected the components and applied teflon tape, which made the system water tight. We then applied soapy water around the compression fittings, tubes, and leaf to ensure that no air bubbles were present during operation. After careful analysis, we found that there were no air leaks anywhere in the system. We also confirmed this during operation as any air still present in the system would exit from the mesh surface, shown in Figure 19 below.
Figure 19: Three-piece Leaf
The tube height, tube width, plate diameter, and number of tubes were easily determined using calipers and tape measures. Each of the measurements either met the specifications or were within an acceptable range.
Figure 20: Charging System
Figure 21: Solar Concentrator
The rate of transpiration was unable to be determined because the nanoporous layer was unable to hold the water. Figure 22 below shows a microscopic image of the mesh and what we believed to be the nanoporous deposition. After we found that the water would not hold, we took microscopic images of our own to see what may have occurred. Figure 23 shows an image of the multilayered mesh. The manufacturer data of the mesh showed a 5-ply mesh that went from thick to thin layers. However, our imaging showed that there was a thicker layer of mesh on top of the thin layer below. This led us to conclude that ORNL was unable to consistently deposit the nanoporous layer on top of the thinnest layer. The thickest layer caught portions of the nanoporous deposition, preventing a continuous layer of nanopores along the finer layer. This means that no matter how many tests were performed, the water would not be able to be maintained in the pores.
Figure 22: ORNL Lab Microscopy
Figure 23: Microscopic Imaging of Mesh Surface
Conclusion
The year largely consisted of design work, and culminated in a nearly fully functional final product. We generated customer needs and target specifications based on the requirements provided by our faculty advisor. Next, we generated initial concepts and performed down selection to ensure that the generated specifications were properly met. Then, we performed detailed analysis utilizing the driving equations of the overall system and generated CAD of the sub systems. After finalizing the CAD of the system and locating specific part numbers and prices of components, we ordered the necessary parts. Finally, we constructed and attempted to test our system. All components and subsystems were functional except for the nanoporous mesh, and we determined the most likely cause for the failure of this component. Our work this year should provide future teams with a platform from which to test a new nanoporous mesh and expand the design to generate hydropower. We have discussed potential additions, such as a solar still that would condense water vapor from the leaf and transport the water back downwards over a turbine to generate electricity.