New Equipment design delivers accurate coal measurement
The article, “New equipment design delivers accurate coal measurement,” in the February 1997 issue caught my attention because I am involved in research associated with coal and its sampling, analysis and handling. the authors of the article suggest a reasonably simple solution to the `dry basis` feeding coal. A simple, inexpensive and highly reliable method for accomplishing this task could bring the coal-fired utilities one step closer to calorimetric feeding of fuel to boilers yielding improved control of the fuel-air ratio, lower excess air, lower corrosion rates to the inside of the furnace and reduced NOX emissions.
The authors offered data to support their thesis, and this is where I run into trouble with their argument. Aside from few minor math errors, their conclusions regarding constant dry-weight feed are based on statistics derived from only three useful test run: Numbers 2, 3 and 5. In addition, the tests were performed on coal samples with virtually NO moisture variation. In my experience, sampling bulk materials to obtain samples that are representatives of the gross lot is difficult, and sampling coal is especially troublesome. The .32 percent variation in moisture reported among all of the tests nearly falls within the range of expected error observed by most laboratories performing analysis of several aliquots taken from the same coal sample (reference ASTM D2961-87, paragraph 9.1). It is not surprising then, that no variation in dry weight was observed between the three tests–there wasn`t any. I don`t understand the application of statistics to three pieces of data and I don`t accept the conclusion.
Even if there had been a sufficient number of tests performed against which a statistical approach was meaningful, reporting a single sigma error is poor practice. Most equipment suppliers report performance variability in terms of two sigmas or COV (the authors` `variation`). Two sigmas is not as attractive of a result but it is more representative of the performance of the process and thereby, less misleading.
The bar graph of Table 2 presents data from tests conducted on Gadsden coal at the suppliers` testing facility. This time the moisture is reported to range between 6.7 percent and 11.5 percent–a 4.8 percent spread. While the table appears to report a constant dry-weight feed, it is offered without detail or explanation of the tests or data that can be scrutinized. Since any coals exhibit a parabolic-like bulk-density curve with respect to changes in moisture, is it possible that the chosen test range coincided with the flattest portion of the response curve?
It is possible that a new method can be developed that can deliver coal on a dry basis, but it wasn`t demonstrated in this article, and cannot be without a statistically backed demonstration that the device can perform with reasonable variations in moisture, consist and coal type.
needs a name, company?????