Optimisation of Cement Kiln Flames

Tom Lowes – OPECS UK, Joana Bretz – CINAR Brazil,
Lawrie Evans – JAMCEM Consulting

 Optimum Flame Theory (OFT)

There is a lot of myth and folk law associated with flames in a cement kiln: which is best, long and floppy or short and hard? Burner on the nose ring, in or out of the kiln? How about burner alignment, up the axis or pointed towards the load? Which rules for particle size of coal, petcoke and alternative fuels should be followed? The answer to all these lies in the fundamentals of combustion aerodynamics and the basic law of Jet Mixing, developed by Thring in the 50’s and extended by Spalding and Hinze to put numbers to it. This paper gives fundamentally based guidelines to how to optimise the flame, that has been applied and is still be used on all 5 continents since it was first developed by Lowes and Evans at Barnstone and then applied – in Blue Circle Industries - in the late 70’s, supported by more recent work on MI-CFD by CINAR. The essentials were published by Lowes and Evans in the Inst of Energy Journal in 1989.

The best starting point is the mnemonic SADAM, the following of which will ensure no chemical reduction and a minimum of SO3 driven build ups problems.

S – Size: for coal and petcoke the residue at 90µm is not what many people quote as a rule – 50% of the volatiles, as that was taken over from the water tube boiler industry – this can be up to the same level as the volatiles in particular for a big kiln and or a high volatile coal. For AF (alternative fuels), the material should be essentially 2 dimensional with 100% passing 30 mm to allow high TSR (Thermal Substitution Rate).

A – Alignment: the burner should be located on the nose ring and pointed up the kiln axis. If there is a coating issue at the front end of the kiln, it can be pushed in by up to 0.5m. In cases that for some reason a too long a flame is observed, the burner can be pulled out by up to 0.5m. The only exception to this is when there is a misalignment of the cooler / kiln hood, which distorts the flow along the secondary air flow around the burner. Then adjustment needs to be made to get the best-looking flame, however these situations – which are not frequent - are best optimised via MI-CFD (Mineral Interactive Computational Fluid Dynamics – developed by Cinar Ltd).

D – Dryness: for coal and petcoke, the as fired H2O content should be less than 1% and for alternative fuels, the maximum limit is 15%. For higher levels of H2O, once the input goes above 0.02 kg of water/kg of clinker, the AF should be located above the burner mid-way to the top of the kiln and even pulled back into the kiln hood – for a non calciner kiln, by up to 1 metre.

A – Air: CEMFLAME showed that for coal flames with an optimum momentum the CO began to increase when oxygen at kiln back end is around 1.5% while with petcoke this occurred some 1% O2 higher. Hence for a good kiln operation to avoid clinker and process issues the minimum oxygen concentration at kiln back end (KBE) needs to be set according to these rules.

M – This is Momentum and is the key to all cement kiln flames i.e. Optimum Flame Theory. The optimum momentum for each burner is a function of three main parameters: a) the Kiln oxygen; b) the secondary air temperature; and c) the burning zone thermal load. CEMFLAME in the 90’s was seeking to produce a practical report for its Consortium Members and specified a specific minimum momentum of 7 N/MW for a good flame. This number was adopted as a reference by most of the cement producers and plant suppliers. The main problem with it is that it led many plants down to the wrong path because the N/MW is not a scaling criteria, it is actually a function of the burning zone thermal load. The 7 N/MW worked pretty well at the IFRF (International Flame Research Foundation) for CEMFLAME but only it had only 3 MW/m2 thermal load.

Fig 1 shows how a flame jet with a non-swirling flow expanding at a 9 degrees half angle expands in a kiln. In the first kiln meters, for a distance of three internal kiln diameters, the mixing of the secondary air into the expanding flame is controlled by the momentum of the burner. After that, we have the plug flow region where the mixing of the fuel and oxygen is at the mercy of slow turbulent diffusion. An optimum flame needs to have first the secondary air mixing over the expanding jet boundary and then inside the jet onto the flame’s axis before the beginning of the plug flow zone. Otherwise, longer and floppier flames are produced, with consequences on kiln output, build up and clinker quality.

Flame combustion aerodynamics.png
Fig. 1. Flame combustion aerodynamics

The simple jet mixing equation can be used to quantify what momentum is needed to achieve this optimised mixing:

Mj/Mo = K*x/do

This, after about 4 pages of mathematics for a cement kiln ends as:

G opt = {4 (Ms+Mp) [ 1+0.68 (Tr/Ts)1/2 ]2 }/(p rs D2 )

Simplifying for 900oC secondary air and 2% O2 at Kiln and developing it to use N/MW the equation becomes:

N/MW = 2.9*MW/D2

Where D is the inner brick diameter (m) of the burning zone and MW is the energy going into the kiln.

It clearly shows that the N/MW needed to achieve an optimum mixing/flame is a function of the burning zone thermal load in MW/m2. For example: at 2% O2 and a Secondary Air temperature of 1000oC with 5 MW/m2 a specific momentum of 11 N/MW is required for the optimum mixing.

The use of swirl has been developed by burner suppliers for use in cement kilns based on its success in the power industry. However, swirl application needs to be carefully analysed as it expands the jet (see fig 1). Unless it has benefits of faster combustion, it will cause impingent of unburnt fuel on the stable coating and clinker with consequences on refractory life, clinker quality and SO3 cycles and build up. This is evidenced when most burners with little or no swirl perform better for petcoke and AFR burning.

The IFRF has shown that the best use of swirl is to apply it to generate a swirl induced internal reverse flow zone which brings back hot gases and heats up the fuel as the root of the flame. To produce such an internal reverse flow zone a minimum of 30 degree vane angle is needed. This swirl itself would expand the flame shape too fast and cause flame impingement, so the swirl needs to be able to be contained by a powerful axial flow to get the benefit of the reverse flow zone and not the impingent downsides. Hence a burner that combines swirl and axial in the same outlet nozzles cannot use its swirl effectively. In the same way a burner that has swirl and axial in two separate channels but cannot independently control both the axial and swirl flows, also cannot use the swirl effectively. A swirl angle less than 30 degrees has little value as it does not generate a swirl induced reverse flow zone.

Practical Application

The SADAM mnemonic is a good starting point. Care needs to be taken in evaluating the momentum, as pressure can be misleading. If the burner accelerates its axial or swirling flow before the final exit, which has a larger area, the flow will have expanded to fill a larger area and will have a lower momentum. Examples of this were found in the 2000’s which were resulting to poor clinker quality and short burning zone refractory life.

A useful equation to calculate the momentum if the air flow is known and the exit areas are not available is:

Pressure (mbar) = 0.005*density (kg/m3)*V(m/s)2

This is however only a useful starting point. A complete understanding of the problems perceived by the Plant is also needed, plus a check on the clinker for reducing conditions and insoluble alkalies, which indicates either fuel impingement, CO reduction or over burning.

A check of the KBE CO and NOx versus KBE O2 is a very useful guideline into how the burner’s flame is performing. The CO for a coal-fired flame should not begin to increase before 1.5% KBE O2 and the NOx should peak at  2% O2. Figs 2 and 3 shows examples of good and bad burner operations: the good with OFT momentum and the bad generally with approximately 30 to 50% lower momentum.

Kiln Back and analysis - CO vs oxygen OFT and Low momentum burners.png
Fig. 2 Kiln Back and analysis - CO vs oxygen OFT and Low momentum burners

Kiln back end analysis NOx vs Oxygen.png
Fig. 3 Kiln back end analysis NOx vs Oxygen: OFT and Low momentum burners


The CO is formed when there is not enough secondary air mixed in with the fuel when it volatilises to form OH radicals. The reason for the NOx/O2 peak is that initially the extra O2 cleans up the reducing conditions, but after that, because with OFT the mixing is good, the additional O2 drops the NOx due to a lower flame temperature.

Another very useful check is a sulphur balance and volatilisation study including their driver’s identification. Generally the VF SO3 (sulphur volatilisation factor), which is the ratio of HM SO3/Clk SO3 (sulphur content in hot meal / sulphur content in clinker) should be less than 3. Additionally, if not all the SO3 going in is coming out at 2.5% KBEO2 on a daily basis, it means that reducing conditions are taking place. Then its causes need to be found and eliminated.

This approach has been effectively applied over the last 40 years to more than 100 installations with the aim at low cost modifications to existing burners. However in the risk averse situation that has arisen in the 2000’s, plus many alternative “expert” views extra insurance is needed by a plant before making recommended burner modifications.

This has been provided as needed on many occasions since the early 2000’s by CINAR MI-CFD. Fig 4 shows how the oxygen profile is stratified for a burner operating at 6 N/MW. The fuel burnout is very bad even for all particles below 45µm size. By increasing the momentum of this burner to 10 N/MW – that calculated from the OFT equation –the O2 stratification is successfully eliminated, excellent fuel burnout even for bigger particles such 90µm and a drop of 60oC is observed at kiln back end temperature. It should be noted that oxygen stratification not only impacts on the kiln flame but also on the burnout of feed shelf fuel, on the combustion of unburnt material coming in the hot meal from the calciner as well as making the calciner fuel more difficult to burn.

Fig. 4 Kiln Burner Optimisation with MI-CFD by Cinar Ltd.png

Fig. 4 Kiln Burner Optimisation with MI-CFD by Cinar Ltd

Recent Case Study

When to comes to the application of OFT it is best done in the hands of a person who fully understands the cement making process and the fundamentals of SO3 cycles and build up, how unburnts in the hot meal drive SO3 cycles and the role of Cl in it, as well as the chemistry of clinker and the impact of the flame on it. How to spot reducing conditions when the clinker has no brown centres and the role of ortho rhombic C3A from XRD is also helping diagnose problems with a burner.

The case study in this paper was for a preheater plant that was trying to use 100% petcoke in the kiln with 15% shredded tyres at the into the kiln back end and was have to stop every 3 weeks to dig out the riser and the kiln from build-up.

The first step in such a problem is to look at the SO3 cycles and KBE O2 and CO. The best way to look at the SO3 cycle is to calculate the volatilisation factor for SO3 which is hot meal (HM) SO3/Clk SO3. For this factor the target is 2 and up to 3 is permissible. For this Plant it was > 10. There was no KBE Probe but in the preheater the reading was at 1 – 2.5 % O2 and CO up to 10,000ppm.

A SO3 balance showed that the SO3 going in was not coming out. A 48-hour test was carried out where the HM sample was taken 30 mins before the corresponding clinker sample and the O2 average calculated for the 10 minutes before the HM sample was taken. The results are shown in figs 5 with the clinker SO3 and HM VF vs PH O2.

A SO3 balance showed that the SO3 going in was not coming out.png

Fig. 5 Dependence of SO3 content in raw meal at the entrance to the furnace (a) and in clinker (b) on the concentration of O2 in the baked heat exchanger (PH Plant 65% Petcoke coal mix via Burner and 15 % Tyre shreds on Feedshelf)

These show that the SO3 in not getting out in the clinker and the HM SO3 is was above the target of at 2.5 and both have a strong dependency on preheater (PH) O2

An analysis of the kiln burner was made which showed that the primary air fan could give at 4 N/MW and from OFT the momentum needed was 9 N/MW. Under these circumstances the situation shown in fig 4 will apply in that the petcoke/coal burnout will have been slow and even allowed unburnt petcoke to impinge on the burning zone – driving the SO3 cycles – plus more importantly the O2 at the bottom of the kiln in the region of the tyres shreds will have been very low and hence the shreds would not have been burning out before the end of the calcination zone – 900C in kiln feed. Under these circumstance the rubber in the form of C will react with the excess SO3 over alkalies in the form CaSO4 and send the SO3 into an increased SO3 cycle causing high VF SO3 and HM SO3.

The recommendation to the Plant was to increase the size of the PA (primary air) fan and make the PA holes big enough to take 9 N/MW and run at a PH O2 that keeps the VF SO3 < 3 and get all the SO3 out into the clinker. Fig 6 is the Plant feedback, where they were able to run at 100% petcoke plus 15% tyres shreds on the feedshelf and never need to stop to dig out the kiln anymore

hot meal so3 vs ph o2.png
Fig. 6 Furnace operation indicators before and after optimization of the fuel combustion mode

Currently OFT is being applied to plant problems at 6 different locations as part of an overall look – including MI-CFD in some case – at the process, emissions and clinker quality while aiming for 100% calciner TSR (total alternative fuel substitution ratio) and / or NOx < 500 mg/Nm3 at 10% O2 without SNCR


The use of materials published on the site is allowed only with the reference to the source (the journal «Cement and its application») and a hyperlink to the quoted material.
Share:  
The use of materials published on the site is allowed only with reference to the source (the journal «Cement and its application») and a hyperlink to the quoted material.
© 2007-2024 PetroCem Ltd
Privacy policy