A correlation is a measure of the relationship between two or more parameters. In our line of work, this typically refers to the relationship between the amount of particulate in a pipe or duct and the subsequent triboelectric signal produced.
A triboelectric monitor produces a pico amp signal caused by the interaction of the electrically isolated probe and the particulate in the process. In applications where the process conditions remain relatively constant, the pico amp signal becomes a relative indication of the amount of particulate in the process. For a majority of applications, that is all that is needed, using the pico amp signal or a percentage as the indicator of the particulate level. However, in some cases the user prefers or is required to provide a signal by using actual engineering units such as milligrams per meter cubed (mg/m3 ) or grains per dry standard cubic foot (gr/dscf). When this occurs a correlation needs to be established.
In order to perform a meaningful correlation, we must assume several factors will remain fairly consistent: specifically velocity, material type, and particle size. A correlation between the pico-Amp (pA) signal from our electronics and one material will not be relevant to that of a different material. It is also important to note that the triboelectric sensor needs to be installed in the common outlet after your baghouse, in a location with several straight duct diameters both up- and down-stream. Also confirm that the sensor is not in an open stack, or otherwise exposed to ambient weather conditions (such as rain or snow), and that the probe is situated near the middle of the flow stream, reaching at least halfway across the duct.
It’s best to perform routine wiping-down of the probe prior to performing any controlled testing such as a correlation test. Establish a reasonable baseline (pA) signal at the Auburn triboelectric monitoring device (equipped with either 4-20mA output, or digital communication such as Ethernet or RS-485). Note: with 4-20mA output, if your baseline is set too low, you may not have sufficient resolution to extrapolate good data from the results. Conversely if it is set too high, your signal may be cut off at 20mA and you will lose valuable data. Auburn recommends finding the right balance in the 4-20mA output to get a baseline signal around 10% of scale, yet still keeping your cleaning pulses within the 4-20mA range. Adjust as necessary to your particular application, and be sure that your data collection is logging a data point at least once per second.
Methods of testing for correlation
Once the unit is running normally and data are being collected, a stack test will need to be performed. This is a controlled measurement (collection and weighing) of the actual material coming out of a stack over a specified time period. Often the protocol for a US EPA Method 5 test is followed. However it is usually recommended that at least three samplings be made to provide better correlation. If the emissions levels can be varied this can also help to improve the accuracy/consistency of the results.
With stack testing data in-hand and corresponding triboelectric (pA) signals to match up to that data, you’re ready to make a correlation. At this point you’ll need to plot the pico-Amp signals vs. the time period of the stack test, ideally finding a fairly straight line. If need-be, the data can be averaged to smooth out any cleaning spikes that appear during the stack testing period.
Supposing the data averages out to a constant 500pA over the course of an hour, and the stack test yielded an average of .0001 gr/dscf (during that same hour). You can then extrapolate that, averaged out over time, you are emitting about .0000002 gr/dscf per pA, per hour.
Correlations can only be done on-site, keeping a consistent velocity. As noted above, there are a lot of different factors which will affect the resulting triboelectric signal. As such, it would be impossible for all of these factors to be reproduced anywhere besides in the actual process stream.
Note: The term correlation should not be confused with a calibration, which is a validation of the precision of an instrument. More on this in a future blog post…
Coming soon: establishing a good correlation when your velocity can vary! For more helpful information on how to choose the appropriate Bag Leak Detection System, check out 15 Tips for Selecting the Best Bag Leak Detection System: