Vol.5 No.4 2013

Research paper : Evaluating Uncertainty for the Standardization of Single Cell/Stack Power Generation Performance Tests for SOFC (A. Momma et al.)−259−Synthesiology - English edition Vol.5 No.4 (2013) rated power test on a 100 cm2 planar SOFC single cell. For the test we used the measurement and control instruments that we normally used for performance tests. When there is a large variation in the measurement values, the level of uncertainty may be reduced by making repeated runs of measurements. However, as long as the system is in a stable state, a single run of about 30 measurements (1 measurement/second) would show that the impact of the variation of measurement values on uncertainty is small. The objective of the rated power test is to produce power output, which is obtained by multiplying the current, the input quantity, and the voltage, the measurand. Thus the column on the right end of Table 2 shows the uncertainty converted to power. A comparison of the uncertainty values demonstrates that the majority of uncertainty derives from the temperature measurement of the cell. It is also clear from the breakdown of this source of uncertainty that the thermocouple used in the test accounts for a significant portion of that uncertainty. One of the benefits of uncertainty analysis is that by producing a data set like Table 2, one can easily see which measurement needs to be improved in order to reduce the level of uncertainty in the measurand. By changing the thermocouple used for the measurement from a type-k thermocouple (class 2), which was used for this measurement, to a class-1 thermocouple, it is possible to reduce the standard uncertainty of power output measurement from 0.13 W (0.52 %) to 0.08 W (0.33 %).From the above, we believe that the impact of the uncertainty on the variation of the measured values would remain small, so long as the measurement is made by carefully making sure that the system is in a stable state. It follows, then, that it is possible to derive a rough estimate of the uncertainty of test results solely from the uncertainty of the measuring instruments used in that test. Thus, when the uncertainty of the above-mentioned cell was calculated with the assumption that the measurement instruments used for the calculation had the maximum allowable range of the instrument uncertainty set by the draft standard, the relative expanded uncertainty was estimated to be 1.4 %. Since the draft standard prescribes that an instrument falling within the allowable range of uncertainty be used, it follows that the test operator would be able to obtain a measurement result with the relative expanded uncertainty of approximately 1.4 % at the maximum. However, it should be noted that this value may change significantly from one test unit to another, such as when the unit has a completely different sensitivity coefficient or when the rated condition is characterized by a high fuel utilization rate. 6 SummaryThe maximum variations of input quantities and the uncertainty values of the instruments established in the draft standard were based on a set of working values that were initially prepared by our group and were subsequently discussed and adopted by the committee as the final values acceptable to all manufacturers. In conclusion, our study demonstrated that by conducting a test in accordance with the draft standard, it is possible to obtain a measurement with a relative uncertainty of approximately 1.4 % at the maximum. We believe that the values proposed in the draft standard were reasonable for the commercial transactions of SOFC. In the end, however, the uncertainty analysis and its equation for the rated power test that we proposed were omitted from the final version of the draft standard by the national committee responsible for preparing the New Work Item Proposal (NP) for submission to IEC. As a result, only the maximum allowable variation of the input quantities (control parameters) and the instrument uncertainty remained in the NP. Nevertheless, it was still fortunate that we were able to verify the maximum uncertainty value as described above. The main reason that the uncertainty analysis was removed from the draft standard is that many manufacturers are still not familiar with the concept of uncertainty and they focused above all on the troublesome nature of the uncertainty analysis. In proposing the equation for uncertainty evaluation, we did our best to present an equation (method) that would minimize the burden on the test operator, but we failed to provide sufficient evidence on the merit of conducting uncertainty evaluation. The concept of uncertainty ensures the reliability of data across borders through the traceability system. It is recommended by many international standard organizations and is likely to be incorporated into an increasing number of international standards in the decades to come. At the same time, however, it may take some time to convince the manufacturers in this field that the potential benefit of uncertainty evaluation makes it worthwhile to incorporate it in their testing procedures if they aim to produce quality products.This draft standard is currently being reviewed by an international Working Group (WG), and it may see an introduction of the uncertainty evaluation any time as a result of a proposal by another country. In addition, as far as our national committee is concerned, it appears that the concept of uncertainty is gradually gaining in prevalence among the committee members as a result of our presentation and the discussions held on the topic. Therefore, given that uncertainty evaluation may be proposed at any time, we believe that our work has helped to lay the groundwork for determining whether or not it should be adopted. While we, the authors, are not experts on uncertainty, there


page 35