The developed model was used to simulate the power and cooling for scenario cases with and without storage

The overall test duration of test for each cell was 47 days. Each cell was subjected to steady state testing for 1000 hours at 0.5A/cm2 as shown in Figure 108. Figure 109 show the temperature change during these tests. The cell operating on H2-N2 failed at 850hr. A sudden increase in cell temperature was likely the cause of failure. Consequently, the degradation rate for all cells is calculated based on voltage data from time 0 to 850hr. The degradation rate is calculated based on the slope of best linear fit to voltage over time with R2 value of 98%, 97%, and 93% for H2, N2-H2, and NH3 respectively. The degradation rate observed for H2 case is 2.51%kh-1, for NH3 case is 1.04%kh-1 and for N2-H2 case is 2.6%kh-1. The cell tested with ammonia shows almost 60% less voltage degradation over time than the other two cases. The inductance and capacitance corrected Nyquist plot of the EIS spectra for different flows at the beginning of the test and after 800hr at 1.256A are shown in Figure 111 and , respectively. The residuals from KK tests were between ±1%. The results show a constant ohmic resistance at the beginning of the test however after 800hr test N2-H2 and H2 case have slightly higher ohmic resistance,grow tray while polarization resistance is the highest for ammonia case. The ohmic and polarization resistance are presented at Table 24. It is evident that the increase in total resistance after 800hr test is due to the increase in ohmic resistance.

The similarity in EIS between NH3 and N2-H2 proves the two-step utilization of Ammonia as found in literature. The inductance and capacitance corrected Bode plots and shows that the NH3 and H2-N2 have similar trends compared to H2. The charge transfer region had lower resistance for H2 while the gas diffusion resistance was higher for H2 than the NH3 and H2-N2 cases. This is likely due to higher hydrogen partial pressure in pure hydrogen case, since mass transport losses through the thick anode support may contribute to concentration polarization. These results are consistent with findings in literature . SEM and EDX investigations were carried out on the anode surface to evaluate the effect of operating NH3, N2-H2, and H2 on SOFC cell. Figure 113 shows the SEM image of an anode after being purges with N2-H2 and reduced which is used as the reference cell. Figure 113 and shows the anode micro-structure inlet and outlet after operating on N2-H2. Please note that this cell failed after 850hr of operation therefore the SEM images are taken after 850hr of operation while for NH3 and H2 cases the images are taken after 1000hr of operation. As it can be seen, there is no significant changes to micro-structure of anode inlet and outlet compared to reference cell. Looking at the SEM image of gas outlet region for H2 performed cell signs of potential enlargement of Ni particle is apparent at the surface level which can be starting for nickel migration; however, the effect is not magnificent enough to effect on quantifications Table 25.

The anode micro-structures are nearly the same without any crack for NH3 case as it is shown in Figure 113 and . Further, the nickel particles are not enlarged suggesting that no significant degradation of the cell anode micro-structure occurs after the ammonia exposure for over 1000h operation time. The findings are consistent with findings. Although no remarkable microstructure degradation signs were observed at SEM imaging, analysis of cross section SEM/EDX image is necessary for further investigation of degradation.and the data of the element atomic percentages. EDX quantification was done based on three different areas on each sample and the average is reported in Table 25. The nickel, oxygen, and zirconium atomic percentages after the tests with H2-N2 and NH3 did not change significantly compared to the reference cell. The decrease of the nickel percentage is less than 5% as compared to the reference cell. Furthermore, it seems that there is no formation of Ni3N on the anode surfaces for the 1000h operation time. For N2-H2 case there is a significant lower atomic percentage of Ni and higher percentage of Zr . The changes are due to the presence of ZrO2 agglomeration on the surface of anode as shown in Figure 114 which came from the sintering process on zirconia refractories and, as such, are not relevant to degradation processes. In this chapter degradation of SOFC was analyzed in various conditions. First the SOFC system degradation was evaluated while operating under steady states and dynamic loads. In order to do so, BlueGEN system was installed at APEP laboratory, and the system first operated under steady state at nominal condition for over 4000hr, and then operated under ramp up and ramp down dynamic load for 1000hr.

Results showed that the BlueGEN system degraded 0.59%kh-1 under dynamic operation while depredating 0.025%kh-1 under steady state load. In the second part, the degradation of single cell SOFC directly fed with ammonia was analyzed and compared with externally reformed ammonia and pure hydrogen . The voltage degradation rate observed for H2 case was 2.51%kh-1, for NH3 case was 1.04%kh-1 and for N2-H2 case was 2.6%kh-1. Also, the SEM investigation showed no sign of degradation on the cell anode microstructure occurs after the ammonia exposure.The goal of this research was to first evaluate the integration of an SOFC system with liquid desiccant air conditioning for efficient and reliable data center power and cooling. The second goal was to evaluate the durability of the proposed technology under different dynamic and steady state operation and with different fuels. The goals were achieved by 1. extensive literature review on SOFC technology, degradation, and fuels as well as data center power and cooling and liquid desiccant technology, 2. Developing a dynamic model for BlueGEN SOFC based on EAGERS and validating the model with experimental results, 3. Developing a physical model for liquid desiccant air conditioning and validating the results with literature data, 4. Calculating data center cooling demand based upon cooling type and weather conditions in various locations, 5. Analysis of the integrated system for different weather conditions and determining the storage capacity to meet the annual demand of data centers, 6. experimental set up and long-term test and evaluation of the SOFC system degradation under steady-state and dynamic operation and SOFC cell degradation operating on different fuels. Findings showed that integrated SOFC-AC system coupled with seasonal storage is a great candidate for powering data centers and keeping the server racks in a safe range of temperature and humidity in hot and humid regions. Also, the SOFC technology used in this integrated system can be designed for robust dynamic operation over many years and can likely be fed by ammonia without any negative effect on durability of the system. A spatially and temporally resolved physical SOFC system model and a corresponding physical model for liquid desiccant dehumidification were developed and successfully verified by comparison with experimental data and literature data respectively. To simulate the integrated system behavior to meet server power and cooling demand in a data center,hydroponic trays first the data center demand was modeled based on weather data. The SOFC system could produce enough power for servers in all cases. For a single server case with no storage, the results showed that the SOFC exhaust heat was able to continuously provide almost 25% of server rack required supply air in the safe range of temperature and dehumidification under these conditions. This result showed that to provide enough cooling required for data centers seasonal storage is required.

Results from simulating integrated SOFC-AC with seasonal storage showed that for three of the studied locations the integrated system can provide enough cold and dehumidified air throughout the entire year, and for the extreme weather condition in Texas it meets up to 85% of the demand in a row level design. The analysis showed that the row level design gives the opportunity to run the fuel cell at partial load with higher efficiency and decreases the infrastructure needed for server cooling. Providing the required cooling for servers through liquid desiccant dehumidification using the high-quality heat of the SOFC exhaust will significantly decrease data center electric power consumption. This configuration offers the potential for high energy efficiency , environmental and economic benefits to data center applications, reducing both greenhouse gas and criteria pollutant emissions.First, the BlueGEN system was successfully installed and tested for observed degradation under both steady-state and dynamic operating conditions. To evaluate the degradation of BlueGEN system first the system ran under nominal conditions and then under dynamic load cycling conditions for 4000hr and 1000hr respectively. Findings showed that the system degrades twice as fast compared to steady-state operation while operating at a highly dynamic condition. As the system degrades, the fuel consumption increases and as a result the efficiency drops. Even though the SOFC system degrades twice faster in dynamic operation mode, the degradation rate is within the range of SOFC stack healthy lifetime which makes the BlueGEN SOFC technology a great candidate for large scale installation at data centers. Secondly, the objective was to evaluate the influence that ammonia fuel has on anode supported Ni/YSZ single cell compared to externally decomposed ammonia and pure hydrogen cases. Results showed that the cell voltage and power density were similar for all three cases and the voltage degradation was lower for ammonia compared to the other two cases. No changes were observed on anode surface for ammonia fed cell within the scale of SEM observation. EDX quantification showed no significant changes in atomic percentage of Ni meaning that there were no signs of nickel nitrite formation. Findings suggests that ammonia has a great potential as a substitute fuel for natural gas or H2 in SOFCs. Cultivation of the cannabis plant has been traced to around 10000 BC with the first evidence of the plant being found in China. It is a plant long thought to have therapeutic properties and was used by the Chinese for medicinal purposes. Indians also attributed therapeutic properties to cannabis and greatly used it, dating back to 1000 BC, as an antispasmodic, appetite stimulant, antiparasitic, anesthetic, anticonvulsant, antibiotic, hypnotic, analgesic, diuretic, digestive, and aphrodisiac among even more other treatments.The Shengnung Pen-Ts’ao Ching, known as the earliest Chinese materia medica book mentioned this about cannabis: “ma-fen … if taken in excess will produce visions of devils over a long term, it makes one communicate with spirits and lightens one’s body.” This was the first written record of cannabis having psychoactive activities. William B. O’Shaughnessy, a 19th century Irish physician used cannabis as a treatment for tetanus and various other convulsive diseases, introducing the plant to Western medicine. Other physicians in that time period also tested the plant for the treatment of mental disorders. Therefore, interest in the therapeutic properties of cannabis has a long history. However, prohibition and federal legal restriction of the plant have greatly hindered the study of it therapeutic effects. There are over 400 known compounds form the Cannabis sativa plant, with more than 104 of them labeled as cannabinoids. The rest of the compounds include nitrogenous compounds, flavonoids, terpenoids, and other more common plant molecules. Cannabinoids from the Cannabis sativa plant are also known as phytocannabinoids. The cannabinoids act on cannabinoid receptors, part of the endocannabinoid system in the body that regulates the neurotransmitter release in the brain. Of the cannabinoids, the most common are delta-9 tetrahydrocannabinol and cannabidiol . In the cannabis plant, the cannabinoids are mainly found as their carboxylic precursors which can then be decarboxylated in the presence of heat and light. Both CBD and ∆9 -THC are synthesized in the glandular trichomes of the plant. These glandular trichomes are only present in the female plant; therefore, cannabinoids can only be produced from the female Cannabis sativa plant. The cannabis plant is classified into three categories, primarily based on the ∆9 -THC concentration: industrial hemp where ∆9 -THC does not exceed 15 percent dry weight, intermediate type with near equal amounts of both ∆9 -THC and CBD and the ∆9 -THC type also known as the drug type, having primarily ∆9 -THC. Both ∆9 -THC and CBD act on cannabinoid receptors which were first elucidated in the 1980s by scientists William Devane and Allyn Howlett.