In Boiler Bits 7 we discussed the merits of oxygen trim control and arrived at a most exciting conclusion, namely that a mathematical relationship exists between excess air and the oxygen content of the flue gas. If we know the one, the other can be calculated. So once we have established the best excess air level for most efficient combustion of the fuel, our control system simply has to adjust the air-fuel ratio to produce the corresponding oxygen level. And with a modern zirconium oxygen sensor ultimate combustion control has come within reach! 

Well, in theory at least. There is however one major drawback, as well as a few minor ones.

Ideally the oxygen sensor must be located right there where combustion is taking place – in the furnace. But with furnace temperatures soaring to some 1200 ⁰C and higher the zirconium oxygen sensor has little chance of surviving the combustion environment. So logically we move it to a location downstream of the furnace, normally somewhere between the boiler’s flue gas exit and the ID (induced draught) fan where flue gas temperatures are less hostile. And yes, I am specifically referring to pea coal boilers where an ID fan is employed as a rule to move the flue gas towards the stack, whilst maintaining a negative furnace pressure to keep the combustion gases safely within the confines of the furnace and flue gas passages.

Unfortunately negative pressure spells trouble. It causes ingress of ambient air (also referred to as tramp air) wherever the integrity of the furnace or flue gas passages under negative pressure are compromised. Leaking seals, ash port doors and joints, warped covers, etc., all create openings through which tramp air seeps into the flue gas passages. And to crown it all a short fire on the stoker leaves ample room towards its rear end for combustion air to pass through the grate unaffected by the fire. As much as 50% of combustion air can pass through the rear end of the grate if under grate dampers are not set correctly (mostly not set at all!). So potentially there is a real risk of combustion/ambient air and combustion gas mixing right there in the furnace! By the time this air and gas mixture reaches the oxygen sensor it no longer reflects true combustion conditions. For all practical purposes our expensive oxygen sensor is rendered worthless as the system now controls from a distorted reference and one most likely ends up with a control system suppressing air supply to an otherwise perfect fire to achieve set point oxygen levels. The very device which was meant to optimize efficiency now becomes its worst enemy.

I would also like to point out a number of lesser challenges facing an oxygen trim system:

  1. A zirconium oxygen sensor has a limited service life of some 3 to 5 years, depending on the working environment. Even while lying in a bin in the store it deteriorates (although slowly) and may very well be expired by the time it is required to replace the one in service. Users of steam boilers often don’t know they are letting themselves in for high replacement and maintenance cost when installing oxygen trim systems.
  2. The oxygen sensor needs to be cleaned and calibrated from time to time. Special tools and calibration gas is required to carry out calibration. Or it can be contracted out to a service contractor.
  3. Setting up and tuning the oxygen trim control (PID) can be quite challenging, specifically when firing pea coal on a chain grate stoker. Unlike atomized fuels, coal is extremely slow (and often inconsistent) in its response to changes in the air-fuel ratio, which is often required with load fluctuations. For instance if the excess air is increased by say 10% it may take the combustion system up to 10 or 15 minutes before this change reflects to its fullest extent in an elevated flue gas oxygen level. Thus the challenge facing the control system boils down to “synchronizing” the rate of air-fuel ratio adjustment with the rate of response of the fire to adjusted excess air supply. If syncing is not performed accurately oxygen levels continuously overshoot or undershoot the target set point, especially if steam demand fluctuates significantly during a normal production day.

Then in closing:

  1. I firmly believe that oxygen trim systems come to their right in combustion applications using atomized fuels, such as oil, gas and pulverized coal, where responses to changes in excess air are immediate and consistent. It is definitely not ideal for pea coal firing.
  2. The challenges and demands facing users of steam boilers dictate that state of the art combustion control systems be employed. I am often surprised to find that new control systems are still being supplied and installed without incorporating the control, computing and communication power of plc’s. Optimizing combustion efficiency is sufficiently complex to warrant the use of computer technology in the operation and control of steam boilers.
  3. In a recent development we have successfully tested a virtual oxygen sensor in a control system which uses the plc and special control logic to simulate oxygen trim control. Under normal operating conditions it delivers similar efficiency as oxygen trim control, but without the expense of purchasing, maintaining and replacing the otherwise required zirconium flue gas analyser.

This post was compiled by René le Roux for Le Roux Combustion, all rights reserved. Do you want to know more about combustion control systems and combustion optimization? Please contact us for your professional boiler automation, steam system efficiency and coal characterization needs.

Kindly note that our posts do not constitute professional advice and the comments, opinions and conclusions drawn from this post must be evaluated and implemented with discretion by our readers at their own risk.

Leave a Reply