Optimizing Raw Mills Performance ; the Materials WAY

Introduction:

• Raw milling is one of the most important and integral component of cement production process; it is the stage which produces the most important intermediate product i.e. Raw Meal which is critical not only from cost point of view (Fuel / Electricity consumption) but also for ensuring the quality of upstream products like clinker as well as the desired final product i.e. Cement.
• By carefully optimizing and selecting the raw materials available; not only raw mill capacity will be enhanced but also power savings will be realized at the same quality parameters i.e. Residues / Fineness.
• In case of Raw Mix Pile; final corrections are made through corrective hoppers of raw mills; whereas in case of a totally separate raw materials piles; total blend is carried out during the raw milling stage.
• Most of the raw mills used today are Vertical Roller Mills (> 80 %); although some ball mills are still employed; but consume more energy.
• Material residence time in case of vertical roller mill is much less than the ball mill; hence a consistent flow of input materials is required to prevent the mill from tripping due to excessive vibration whereas ball mill is not that sensitive to the inconsistency of input materials feed.
• Grinding is done by a Compression, Shear, Impact & attrition; depending upon the grinding mechanism used.

Optimizing Raw Mill through Raw Materials Selection and Optimizing Chemical and Physical Properties:

From Materials / Quality point of view; raw mills performance is primarily dependent upon

A)    Raw Materials Size

B)    Raw Materials Moisture

C)     Raw Materials Hardness / Grindability

D)    Raw Meal Modulai Setting (LSF, SM & AM)/ Chemical Parameters of Raw Materials

A.  Raw Materials (Raw Mix) Input Size & Product (Raw Meal) Size

• Generally, Roller mills can accommodate high size (up to 80--100 mm) and high moisture (20 %); whereas Ball Mills can normally accommodate (up to 25 mm) & moisture up to 12 %.
• Traditionally vertical roller mills operate with feed around 80—100 mm size but reducing this to lower size has proven beneficial to capacity enhancement in number of plants; with ball mills the finer the raw materials feed size the better will be the mill output / capacity.
• The raw materials input size will also reveal its significance with regard to mill capacity improvement as tested in Bond Work Index Test; however, in actual practice the size reduction has more impact than what the Bond Work Index test will reveal.
• For Example: Using the Bond formula, the increase in raw mill output was predicted to be only 3.5% for a recent modern VSM raw mill for a reduction in feed size from 70 mm to 30 mm. However, the actual output increase obtained was plus 14% or 4 times that which was expected. From several studies the actual output gain has been between 2 and 6 times more than the theoretical prediction.
• If by reducing the feed size to the raw mill increased the output but also increased the raw meal residue, then this may indicate that there are other milling circuit limitations that are stopping you from achieving the maximum potential benefits like separator and fan performance / bottle necks. If the increase in output also results in a higher raw meal residue then this can indicate that the separator is a bottleneck and it may prove cost effective to modernize the separator by, for example, LVT separator conversion technology.

Below is an example of the results of feed size reduction for modern VSM Raw Mill: (LM 36.4 RM)

Parameters Capacity kwh/tonne Mill Vibration

Max 90 mm 240 8.05 5--7

Max 60 mm 248 7.84 5--6

Max 30 mm 262 7.41 4--6

• In the case of an elevator ball mill, the reduction of feed size from 30 mm to 18 mm produced an increase in output of 8%. This corresponded to 4 times the benefit predicted by the Bond formula.
• Mill output and power consumption not only depends upon the input materials but also on the desired fineness (200 / 90 Micron) of the raw meal produced; hence a careful evaluation and target setting of raw meal residues.
• One of the simplest ways to increase any raw mills capacity is to grind coarser as shown by the table below. However, we need to ask ourselves what are the potential benefits versus the risks? We need to consider the nature of the residue and factors such as its free silica content (Quartz) / pure crystalline CaCO3, natural mineralogy of the raw meal (Easy to Burn / Hard to Burn) and the relationship between residue and raw mix combinability temperature.
• In one case a plant was told it should reduce the raw meal residue from 20% to 10% retained on a 90-micron sieve. This was claimed to improve kiln operation. It had no benefit what so ever as the plus 90um residue material was not high in free silica / crystalline CaCO3 & moreover the raw meal by mineralogy was easy to burn. It certainly affected the kiln since this was limited by raw milling capacity and the 20% reduction in raw mill output resulted in minus 20% kiln output!

Some correlations with Raw Meal Residues v/s kwh/tonne and TPH of a VSM Raw Mill

Raw Meal Residue @ 90 Micron % kwh/tonne TPH

5% +90 Micron 10.14 161

7.5% +90 Micron 8.89 183

10% +90 Micron 8.16 200

12% +90 Micron 7.64 214

15% +90 Micron 7.12 229

20% +90 Micron 6.47 252

• Sometimes the expected performance indicated above may not be achieved at, say, the finer residues. This again would suggest that the separator performance may be the limiting factor. At the coarser residues the limits to raw meal production may become transportation capacity issue etc.
• In authors experience at one of the plants using vertical raw mill; the mill capacity was increased whereas power consumption was decreased when 15---30 % of Crushed Limestone was replaced with Powder Limestone (Limestone Dust) bought from the local crushers in the vicinity of plant; the purchased price was less from the company quarried limestone and in addition to that the saving on accounts of capacity enhancement & power reduction was additional as well as extending the life of the plants own quarry / preservation of the reserves.
• Other alternative raw materials may also prove beneficial to raw mill capacity.

B.     Raw Materials Moisture

• Ball mill can accommodate input moisture level up to 12 % (Max); whereas Vertical Roller Mill can accommodate moisture level even upto 20 % Max.
• Although excessively high moisture level in the raw mills derate the mill but still a particular amount of moisture is required to stabilize the bed and to prevent the vibrations in case of Vertical Roller Mill.
• Roller mills are prone to vibration due to due to poor roller table condition, incorrect dam ring height or poor water injection all leading to unstable grinding bed. A major cause of bed instability is fine, dry mill feed, which can usually be mitigated by spraying water directly onto the bed. That is the reason that raw meal target moisture is normally maintained within 0.25---1.0 %; furthermore, max raw meal moisture i.e. < 1.0 % is kept to ensure flowability of raw meal.
• Raw Materials Input moisture is dried by using hot gases from the pyroprocessing; hence the level of moisture in the raw materials is normally a limiting point in designing the number of pyroprocessing stages (5th / 6th stage) as well as designing WHRPP; furthermore, is also related with thermal efficiency of the kiln. Some plants also use dedicated hot gas generators (Although at additional cost of production); especially to be used when kiln is down and production targets have to be met. Furthermore, heat generated from the main drive also helps in raw materials moisture reduction.
• Normally when designing a new cement plant to handle higher moisture raw materials, the gas circuit design should be optimized to avoid the use of a hot gas generator HGG. The proven techniques include directing the cooler exhaust gases to the raw milling circuit and allowing some raw material to bypass one of the cyclone stages to increase that preheater exit gas temperature.
• Dry materials having 2 to 3% moisture uses 5 or 6 stage cyclone precalciner kiln having thermal efficiency of 730 to 700 Kcal/Kg of Clinker whereas extremely wet materials with 17% raw feed moisture may employ only 3 cyclone stages and 850 to 890 Kcal/Kg of Clinker; hence the design / number of preheater stages along with the resultant thermal efficiency of pyroprocessing is predominantly depending upon the moisture level in the raw materials.
• When the moisture level in raw materials is too high; often a higher capacity raw mill fan & large separator / classifier is required.
•  A review of recent raw mill designs indicates that increasing the maximum design raw material moisture from 4% to 17% would result in a doubling of the installed fan capacity. The separator size must also increase to comply with a typical gas velocity value around 4 m/s through the separator cage. 
• It should be appreciated that when optimizing any raw mill, especially VSM ones, making sure that there is adequate Gas flow through the mill separator and fan can be equally important as optimizing the mill grinding efficiency. You could install the best available ball mill or VSM grinding unit but it will never achieve the optimum system efficiency if you don’t have a well sized and efficient separator and fan to go with it.

C.   Materials Hardness / Grindability

• Specific power consumption of a raw mill is primarily dependent upon raw materials hardness and harder the material more will be the power required to grind the material to the desired fineness level at 200 / 90 Micron.
• Mohs index is used to check the Hardness of Raw Materials; whereas HGI & Bond Work Index is used to correlate the Grindability of raw materials.
• From production / process point of view; the common parameter used for assessing raw materials performance is HGI & Bond Work Index.
• Following tests are used to evaluate the Grindability of raw materials.

a)    Hard Grove Index

• Raw Materials having low HGI are difficult to grind; whereas raw materials with high HGI are easy to grind. It may be noted that it exhibits a nonlinear change in difficulty to grinding. For instance, a change in difficulty from 30 to 40 HGI is greater than a change from 60 to 70 HGI. As a simple rule the output of a raw milling plant is inversely proportional to the HGI of its raw material mix and works should carry out HGI tests on the raw mixes and the individual raw material components.

b)   Bond Work Index ----Both Rod Mill (Wr) and Ball Mill (Wb)

• The Bond test is used to predict the energy required for coarse material grinding between the 80% passing feed size and the split size defined as 2.1 mm. This uses the Wr which is the Bond Rod Mill Index in the Bond Formula. This test is often ignored by plant suppliers but is critical to studies into mill feed size reduction.
• Similarly, the energy for fine grinding the split size 2.1 mm and the final raw meal product residue is predicted using the Wb work index.
• Both values are added together to produce the Bond kwh/ton Grindability.
• Once Grindability tests are done; usual approach is to compare the tested kwh/tonne of Raw Meal (i.e. that which you have tested from the Bond Grindability test results using the Bond Formula) with the actual Raw Mill Main drive kwh/tonne of Raw Meal.
• Bond Work Index predicts the Net Power and you have to allow for typically 5% motor and drive losses to get the Gross Motor Input Power.

 Ratio of actual to predicted kwh/tonne are as under for Vertical Roller Mill.

• 0.5 Good performance 0.55 Typical performance 0.6 Safe design value 0.75 poor performance with lots of scope to improve 
• This means that if the mill is in actual field running is giving less kwh/tonne from the Bond test; than that mill is a good mill

Ratio of actual to predicted kwh/tonne are as under for Ball Mill.

• 1.0 is the typical best performance for a ball mill with the optimum internal configuration i.e. an elevator mill with high efficiency separator and fine feed size. In the optimum elevator mill the discharge from the ball mill is conveyed mainly by an elevator and hots gases ex the mill enter the separator gas flow. Provided that the components are well sized using established guidelines then this is the most efficient system.
• 1.25 typical good mill with some scope to improve.
• 1.5 plus can either be poor performance and or may be partly due to design of mill itself.
• The worst known performing ball mills have been found with a ratio of around 1.7 and even as high as 2.8 when there is no effective gas flow circuit.
• For example, an air swept raw mill with either a static separator or first generation separator may be more sensitive to feed size and gas flow than a bucket elevator mill and will not achieve the same efficiency levels.
• The higher the bond factor for the mill (i.e. Actual kwh/tonne over Bond predicted kwh/tonne) means the mill is less efficient.

BALL MILLS ARE NORMALLY ABOVE 1.0 WHILST VSMs ARE LESS THAN 1.0

Ratio between Wr/ Wb:

• The ratio of Wr/Wb is very important as it is typically around 1.34 as a global average. However, the ratio can range from typically 0.8 to 1.9. This figure relates to the Grindability of the raw materials and not to the design or type of raw mill used.
• However, if the Wr/Wb ratio is higher than 1.34 then it can indicate that the benefits of reducing the feed size to the mill can be a very cost-effective means of optimizing the raw mills (whether they are ball mills or VSM)
• As with any process change you have to make sure that when you reduce the mill feed size you do not end up transferring the overall plant bottleneck from the raw mill to the crushing plant

Other raw mill optimization considerations

• Hard materials often result in increase and speedy wear / tear of the rollers and table and with roller and table wear the Vertical Roller Mill output may drop by 10% with an increase in the power consumption of the mill i.e. kwh/tonne of Raw Meal.
• Hence the need to maintain the condition of these wearing parts to maintain efficiency by in situ welding and new wear resistant materials being developed by several suppliers.

There are several different techniques which have been applied to improving VSM raw mill performance such as

1. Variable speed drive for mill main motor if materials have difficult characteristics in the grinding bed. There is some evidence to suggest that higher table speed may increase raw mill output.
2. Maximizing the separator gas flow and raising the separator speed can give benefits to both raw and cement milling circuits provided that the separator is efficient.
3. Minimize in leaking air.
4. Modernization of separator using LVT technology etc.
5. Replacement of static separators by high efficiency dynamic separators
6. Separate grinding of the more abrasive siliceous raw material in a separate mill such as a sand mill, in case the quantity in the raw mix blend is more.
7. Increased grinding pressure. This is linked to other factors such as the mechanical condition of the table/rollers plus the correct positioning of mill water injection systems.
8. Use of grinding aids – these are common best practice with VCM and Ball Mills for cement grinding but appear to be less commonly used with raw milling. Grinding aids not only help milling efficiency but can also boost separator performance.
9. High Level / Software Control – this can be a useful feature to fine tune any milling system. However, it is not the only solution and we recommend getting the hardware optimized first and then using the HLC for optimizing control and minimizing raw meal residue and raw mix variability.

In the case of optimizing ball mills, please consider grinding aids, reducing in leaking air, HLC, replacement of static separators by dynamic separators as noted above.

In addition, the following options may also be considered for ball mills: -

• Do not rely only on equipment suppliers to sell you what they want to sell when it comes to optimizing raw mill internals and the author has experiences where mill internals had to be modified to avoid loss of output. Common areas where problems have occurred include undersized diaphragm slots, insufficient reverse step lining grinding length, undersized ball size to crush nibs, incorrect internal configuration which limited the maximum ball charge etc.etc!
• It is possible to increase mill output by maximizing separator gas flow/speed etc.
• Optimize the gas flow balance by experimenting with varying the balance between the gas flow through the mill and the booster gas flow and/or separator recycle gas flow.
• Always aim to maximize the mill power drawn by using the optimum charge in the mill. We have seen many claims that running below max motor input kw is less efficient and this theory has been disproved on several installations. Maximizing raw mill output by correct loading also helps to reduce the ancillary power consumption per tonne raw meal produced.

HGI Data for some cement raw materials is as under.

• Limestone 54---78 , Clay 97 , Silica Sand 24---55 , Iron Ore 38

Bond Work Index for Some cement raw materials is as under.

• Limestone 10.2 , Clay 7.1 , Sandstone 11.5 ,Bauxite 9.5 , Gypsum 8.2

Bond Ball Work Index for Some cement raw materials is as under. (FC Bond Paper)

• Limestone 11.30 , Shale 16.40 ,Bauxite 9.45 , Gypsum 8.16, Silica Sand 16.46 , Iron Ore 15.44 , Sand stone 11.53 & Dolomite 11.31

* It may please be noted that the HGI & Bond Work Index for a particular raw materials will vary from region to region depending upon composition / mineralogy.

In house method to determine the raw materials Grindability:

• In case a plant does not have the Bond Mill and HGI apparatus; still the Grindability of raw materials can be tested using following method in a simple and easy way.
• Collect representative sample of raw materials (Minimum 10 Kg), crush to the same size (Say 20 mm) in a laboratory crusher and then take same weight say (05 Kg).
• Pulverize that 05 Kg sample in a common laboratory ball mill for same time (Say 10 Minutes).
• Collect the pulverized sample and check the residue on 200, 90 & 45 Micron for comparing the Grindability of raw materials.

D.  Raw Meal Moduli Setting / Chemical Parameters of Raw Materials (LSF, SM & AM)

• By careful selection and optimizing of Raw Meal Moduli’s; raw materials input %age can be modified resulting into capacity enhancement and power saving but without affecting the quality.
• In siliceous Raw Materials the higher the %age of silica; the difficult it will be to grind as well as more difficult for burning in the kiln especially if it is in the form of uncombined crystalline silica i.e. Quartz.
• Quartz < SilicaSand < Sandstone < Slatestone < Shale < Clay (Easiest)
• Whereas in Calcareous Raw Materials; higher the purity i.e. CaCO3 % Grindability will be hard especially if it is in the form of crystalline uncombined CaCO3 i.e. CaCO3
• Calcite < Marble < Limestone < Marl < Chalk (Easiest)
• In authors experience at one of the plants; Vertical Raw Mill Production capacity increase & power decrease was realized by decreasing the Slatestone % and by Increasing the Clay %; by optimizing (↓) the SM value of Raw Meal. Whereas at other plant having combination of VSM & Ball mills, the production capacity was increased and power consumption was decreased by increasing the Clay % and by Decreasing the Limestone %age; by careful selection and blending of raw materials from quarry.

Conclusion:

Careful selection and optimization of the existing / available raw materials in the vicinity of the plant will not only help to optimize the costs; but will also help to enhance the production capacity at the same good quality raw meal.

There are many ways and means of optimizing raw mill performance and some ideas are indicated above. However, the first essential step is to determine what the actual process bottlenecks are and these vary from plant to plant and there is no single “universal” solution.

PS*:   The author is thankful for the technical input and guidance of Mr. David Baird in the writing of this article.