# Question regarding air flow in a closed circuit

1. Jul 16, 2008

### NightSwimmer

I am new to this forum. I am an applications engineer. I design, build and program industrial automated control systems. I am currently working on a project for a lime calcining kiln. Our project involves an upgrade of the coal-fired kiln's fuel delivery system. This involves a coal grinding mill with a large fan that forces air through the coal mill and carries the coal particulates in suspension into the kiln via an air stream.

We have replaced the fixed speed fan with a variable speed fan. Our objective is to maintain an air mass to fuel mass ratio of 1.8 to 1 in order to maximize fuel efficiency. Our fuel mass is maintained at a relatively constant rate by a closed loop control on a weighfeeder belt feeding coal to the mill through an airlock valve that maintains a seal to atmosphere so that the mill is part of a closed circuit regarding the air flow.

We have installed a hot wire air mass sensor in the return air line from the kiln to the coal mill. It is not feasible to measure the air stream in the outgoing pipe between the coal mill and the kiln because of the suspended coal particulates in that air stream, although this is the region of the air circuit for which we are actually interested in the quantity of the air mass. There is a significant temperature differential between the incoming and outgoing air stream and the incoming pipe is about twice the diameter of the outgoing pipe. The incoming air is approximately 535 degrees Fahrenheit while the outgoing air from the mill is approximately 180 degrees Fahrenheit. Our instrument compensates for this temperature differential and calculates the actual air mass in pounds per hour rather than reporting the volume of air flow in CFM.

I know that Ohm’s law dictates that electrical current in a closed circuit must be the same at all points in the circuit. Does this same rule apply to air mass in a closed circuit? In other words, are we to be assured that the air mass entering the mill is equal to the air mass exiting the mill at a lower temperature? This would imply that the volume of air in the outgoing portion of the circuit must be less than volume of air entering the mill.

We are experiencing a lower reading of air mass than expected. I am concerned that this is being caused by turbulence in the pipe in which our instrument is installed. We were not able to install the sensor in as long a run of straight pipe as would have been ideal because of the required geometry of the piping between the kiln and the coal mill. I have maintained that the mass must be equal at all points in our air circuit, but other engineers involved in the project disagree with my assertion.

The situation is complicated by the fact that the circuit is actually open within the firing end of the kiln and the air stream exiting the coal mill isn’t pure air, but an air/fuel mixture. Am I right or wrong about air mass entering the mill = air mass exiting the mill?

2. Jul 16, 2008

### Staff: Mentor

Yes, in a closed system with a temperature change, mass flow rate is constant and volume flow rate is not. You may want to check to see how much pressure variation there is, but it probably won't be enough to matter. Here are the thermodynamic properties of air: http://www.engineeringtoolbox.com/air-properties-d_156.html

You'll have to convert from SI to English, but I'm guessing the density difference is somewhere on the order of 2:1.

3. Jul 17, 2008

### m.e.t.a.

If a system is closed, the rate of mass flow must be the same at all points in the system -- over a long enough observation period. To me, however, the term "closed system" implies endless recycling, i.e. no mass enters or leaves, but is only circulated. A lime kiln is no such system. You take your inputs (air, fuel, limestone) from outside the system; and when the reaction is over you expel your product (quicklime + impurities) and vent your exhaust gases to the outside. Also, mass flow rate is not necessarily equal to air-mass flow rate.

I am not intimately familiar with the mechanical and chemical processes involved in lime calcination. Could you provide more details, i.e. regarding the type of kiln; the reliability of the sensors used; the % discrepancy between expected and actual air-mass; how you calculated this %, etc.? My (and others') answers will less resemble blind guesses if we can break your problem down into more easily understood pieces.

As far as my blind guess goes, my instinct is that the purity of the input limestone is the biggest unknown. Could the difference between expected air-mass and measured air-mass be as a result of the oxidation (or decomposition etc.) within the kiln of unknown impurities contained in the limestone? If so, air-mass would be removed from, or added to, the kiln exhaust flow as a result. And (conversely) solid-mass would be added to or removed from the quicklime (+ impurities) output.

To truly account for all mass flow in the system, it would be necessary to measure at least all of the following, such as may be relevant:

• The mass of the limestone input (per unit time).
• The chemical composition of the limestone input, at the moment of input. This is needed to calculate the expected mass of gas that a given batch of limestone should yield. Perhaps minor calcination occurs in the limestone before it even enters the pre-heaters (assuming rotary kiln) due to heat leakage. If so, its chemical composition would not be as expected.
• The mass (and, to a small extent, chemical composition) of atmospheric air that enters with the limestone input (per unit time).
• The mass and chemical composition of the coal dust input (per unit time).
• The mass of atmospheric air blown into the kiln with the coal dust by the variable-speed fan (per unit time).
• The mass and chemical composition of exhaust gases that seep out through the limestone's entrance ducts (per unit time).
• The mass and chemical composition of the quicklime output (per unit time).
• The mass and chemical composition of gases (+ unburnt fuel & impurities) exiting through the kiln's exhaust pipe (per unit time)
• The mass of exhaust gases which seep out with the quicklime product (per unit time).
• The accuracy of the air-mass sensors. Also, the conditions under which the air-mass sensors are reliable.
• The temperature at all points in the system; and all sources of heat loss. Temperature dictates the rate of chemical reactions (in parts of the system where reactions occur, i.e. kiln, pre-heaters). Also, the rate of unwanted leakage of gas to the atmosphere (e.g. at the limestone entrance ducts) increases with the pressure of those gases, which depends of temperature. Knowing the temperture enables the calculation (with some uncertainty) of certain of the above factors which are not measurable via instruments.

As the expert, your expert opinion counts a thousand times more than any of my uninformed ones. You're probably right about the turbulence. To find out, as a long shot you could contact the manufacturer of the air-mass sensor and ask them under what conditions their sensors can operate reliably. If you're lucky, turbulence in your pipe will turn out to be the problem and then your proposed solution (locate the sensor in a longer section of the pipe), which was unfortunately unworkable, might instead be replaced with a solution involving the reduction of turbulence in the pipe to a level more tolerable to the sensor.

- m.e.t.a.

4. Jul 17, 2008

### NightSwimmer

First, I would like to thank both of you for your valuable input.

m.e.t.a.,

I'll admit that I did try to ask a simple question regarding a very complex system. Volumes have been written regarding this process, and I just wanted verification that, aside from transient oscillations, the air mass (impure though it may be) entering our coal grinding mill would be equal to the air mass exiting our mill. The combustion airstream is secondary to a much larger airstream created by an induced draft fan that draws air through the entire rotary kiln and subsequently through a baghouse filtration system. The ID fan airstream captures the bulk of the gasses generated by the calcining process along with the majority of the combustion byproducts generated by the firing system. There are also various tempering dampers that introduce cool air from atmosphere into the airstreams in order to maintain optimum operating temperatures throughout the kiln.

As for our expected air mass, we relied on data from Combustion Engineering (the original manufacturers of the mill and fan system) regarding the specifications for the coal mill system as it existed prior to our modifications. While the fuel feed rate has always been variable, the fixed fan speed resulted in a reduced air to fuel ratio at higher feed rates. This can increase the percentage of combustible gasses in the ID fan airstream to unallowable levels. It is also wasteful of fuel (actually a mixture of coal and coke) that becomes ever more expensive. We are able to exceed the previous volume of airflow with our variable speed fan. We can also mimic the old fan exactly by operating the new fan at a particular speed. The air mass readings that we measured both on the kiln with the new fan installed, and subsequently on an identical system on a second kiln that has yet to be upgraded do not coincide with the data provided by CE. We expected to see 14,400 pounds per hour of airflow and we are instead measuring 9,440 PPH.

We have had a manufacturer's representative for our air mass detector on-site during installation, configuration and testing of the instrument. He acknowledged that we were slightly below specification in the length of straight pipe recommended to assure laminar airflow where we are porting our instrument into the return air pipe. He felt that the specs were very conservative and that we should be able to obtain accurate readings with our installation.

I am more concerned with repeatability of the measurement than with the accuracy, since my responsibility is to provide closed loop control. My controls can work well even if there is a consistent and linear error in the air mass readings. The project manager / plant engineer is much more concerned with the accuracy of the measurement than am I. If we can determine that the measurement is incorrect, then we can make configuration changes in the instrument to correct the error. This instrument actually only directly measures the velocity of the airstream inferred from heat loss in the exposed hot wire. We are required to input data regarding the pipe diameter, etc. so that this measurement can be converted to air mass flow information.

5. Jul 17, 2008

### m.e.t.a.

NightSwimmer, thank you for the detailed feedback! I hope I understand you correctly -- or, understand as much as I can. In retrospect, my attempts at "solutions" were slightly ridiculous; however, the discussion is interesting.

The manufacturers of the coal mill and fan system (and the kiln also?), Combustion Engineering, tell you that under a certain set of conditions, X, your exhaust gas flow should reach 14,400 pounds/hr. Your plant meets all of the conditions specified in X, and yet 9,440 lb/hr is measured?

I take it you want to find out what the most likely cause of this discrepancy is; and you want to know if you can justifiably override the installation technician's original calibration settings with settings of your own. Your own calibration would presumably "scale up" 9,440 --> 14,400, either linearly or through some formula. But before you do that, you want to first logically deduce that the exhaust output must be 14,400 and that the sensor must be wrong. (Do I have this correct?)

The only logical possible causes of the clashing data that I can think of are the most obvious ones:
• The data given by CE are erroneous.
• There are problems or unknowns in the manufacturing process (but it sounds like you have ruled these out).
• The air-mass sensor is faulty (this is unlikely considering that an identical sensor in an identical plant reported the same air-mass flow).
• The air-mass sensors in both plants are functioning correctly but are being thrown off by some elusive factor. From your description of the way the air-mass sensor works, it sounds as if the device would need to be told both the temperature and the heat-conductivity of the surrounding fluid (as well as the pipe diameter, etc., as you mentioned) in order to deduce the velocity of that fluid. Assuming that it compensates for temperature fluctuations on-the-fly, it would still need to be fed a figure for heat conductivity. Impurities and unspent fuel in the airflow would alter this conductivity and skew the device's reading. (This might come under "Problems...manufacturing process". Also, this theory falls apart if impurities increase heat conductivity. Perhaps there is a factor, a non-obvious one, which eludes notice and is fooling the sensors?)
• The air-mass sensors in both plants are functioning correctly but both were wrongly calibrated by the technician when they were installed (perhaps the installation technician calibrated both sensors using redundant or incorrect data).

Whatever the true cause of your missing air-mass flow, it is obviously a risky business to take the (almost religious, if you'll forgive the comparison) line of reasoning: "The manufacturers say [...] but we observe [...] therefore our observations must be wrong." I would be more inclined to trust the instruments until such a time as they can be proved faulty!

- m.e.t.a.

6. Aug 3, 2008

### NightSwimmer

We still have not resolved the disparity between our projected air flow readings vs. our actual readings. I am of the opinion that this disparity is due to a combination of erroneous information from CE Raymond and kiln gasses present in our air stream. We still have not ruled out the possibility of air stream turbulence at the point in which we have installed our air mass detector.

I am, however, happy to report that in spite of uncertainty regarding our air flow measurement, we have dramatically improved the fuel efficiency of the kiln firing system. We are generating more heat with less fuel and we have simultaneously reduced the level of combustible gasses in our kiln draft air stream.

I would like to thank you again for your insightful contributions on this forum.

7. Aug 3, 2008

### FredGarvin

There are a couple of things I would double check. The first being the actual pipe dimensions being input. It sounds like this system has been in use for quite some time. You undoubtedly have pipe fouling of some degree. It could alter your pipe ID enough to introduce some decent error.

Second, have you checked your piping system for leaks? Leakage is a big problem in many piping systems.

One last thing...do you have the option to try out other forms of mass flow measurement? Ideally you could get a coriolis meter in there or even an orifice station. I have personally never been a fan of hot wire anemometers. Who knows, you may possibly get a rep to come in and demo a different unit while you are running to do a back to back compare.

8. Aug 4, 2008

### NightSwimmer

All of our piping is new. We are concerned that we might have leakage entering the mill via the rotary airlock valve. Thanks for your input.

"The best engineers are the ones that had the coolest treehouses when they were kids." - NightSwimmer

9. Aug 17, 2008

### NightSwimmer

We have recently discovered that both of the instruments that we received for use in this project are defective. We are currently working with the manufacturer to determine why this happened and to obtain properly functioning mass flow detectors.

10. Aug 18, 2008

### FredGarvin

That's good news. There is nothing worse than chasing your tail looking for phantom leaks and issues when the instrumentation is staring you in the face. I'd be interested to hear what the issues were that made them defective.

11. Aug 20, 2008

### NightSwimmer

We have a factory rep visiting the job-site tomorrow. I hope we'll learn what happened then. Maybe we just got a bad batch of electronic components or maybe there was some sort of shipping damage. It could be something as simple as a botched firmware installation. These instruments have on-board microprocessors.