Coal Flow Aids Reduce Coke Plant Operating Costs and Improve Production Rates

Ron A.Bedard Stelco Inc. Nanticoke, Ontario N0A 1L0 Tel.: 905-527-8335 (extension 2609 or 5609) Fax: 519-587-7709 E-mail: [email protected] David J. Bradacs GE Infrastructure Water & Process Technologies Simcoe, Ontario N3Y 3B2 Tel.: 519-429-2291 E-mail: [email protected] Robin W. Kluck GE Infrastructure Water & Process Technologies 4636 Somerton Road Trevose, PA 19053 Tel.: 215-942-3374 Fax: 215-942-3295 E-mail: [email protected] D. Casey Roe GE Infrastructure Water & Process Technologies 4636 Somerton Road Trevose, PA 19053 Tel.: 215-942-3493 Fax: 215-942-3295 E-mail: [email protected] Ben P. Ventresca GE Infrastructure Water & Process Technologies Stoney Creek, Ontario Tel.: 905-308-4401 E-mail: [email protected] Key Words:Coal, Coal handling, Coal flow, Coal flow aid, Coke, Coke by-products, Material handling, Synthetic fuel, Synfuel INTRODUCTION Over the past few years, North American coke producers have experienced rising coal prices and short supplies of various grades of coking coal. Some prime grades of coking coal are no longer available, stockpiles are being reduced, and it has been

AISTech 2004 Proceedings - Volume I

613

necessary for some coke plants to begin using alternate sources of coal. Synthetic fuels (synfuels) and lower grade coking coals are now being used more frequently throughout the industry. One reason for the tight supply of coking coal is that North American electric power producers have been increasing their consumption of coal while also increasing their stockpiles.1 At the same time, North American coal production rates have been relatively flat, even declining in some areas due to economic, environmental or regulatory pressures. Increases in coal costs and reduced coking coal availability throughout the industry have been accompanied by a rapid rise in the price of coke. Coke prices have been largely influenced by China, where consumption of coke has been surging to feed robust steel production.2 China is by far both the largest producer and consumer of coke, producing nearly 40% of the entire global output. In a move designed to stabilize Chinese domestic coke supply, the Chinese government has recently announced plans to discourage exports of coke and coking coal. The Chinese government hopes this action will offset rampant production capacity expansions which would strain Chinas domestic resources and impact negatively on the environment. In the short run, this move will undoubtedly result in continuing upward pressure on the price of coke throughout the world market. Given that North American coke producers are faced with both higher coke prices and reduced availability of quality raw material, there is much pressure on coke plant operators to maintain high production rates while being forced to use inferior raw material. Many of the currently available coking coals and synfuels are more difficult to handle than prime grade coking coals. These coals tend to agglomerate more readily, causing rat-holing in storage bunkers and pluggage of chutes and hoppers throughout the coal handling system. Delays in filling coke ovens due to poor coal flow result in lost production and reduced overall plant profitability. Frequent coal flow problems throughout the coal handling system result in higher maintenance and labor costs. Improper filling of coke ovens due to coal agglomeration can even result in damage to the coke oven itself. While it may be possible in some cases to mechanically alter or upgrade the coal handling system to accommodate different coals, this would require significant capital investment on the part of a coke producer. In addition, such a venture would most likely require some plant downtime, a difficult decision to make during an economic environment where coke prices are high and continuing to increase. This paper investigates the use of chemical additives specifically designed to improve the coal flow properties of synfuels and low grade coking coals. These chemical additives may allow coke plants to operate more efficiently, while avoiding major capital expenditures to upgrade coal handling systems, by allowing the use of lower grade coking coals which can be difficult to handle. Two case studies are presented which document the performance of the new chemical treatment technology in operating coke plants. BACKGROUND Coal surface moisture can have a significant impact on coke plant coal handling operations. As illustrated in Figure 1, dry coals often cause dusting issues, while wet coals can cause coal handling pluggage problems.

Coal Handling Problems

Dry coal dusting problems

Wet coal pluggage problems

10

12

14

16

18

% Surface Moisture

Figure 1 Coal Handling Properties as a Function of Surface Moisture

614

AISTech 2004 Proceedings - Volume I

For dry coals, fugitive dust emissions are an environmental, health and safety issue. However, mechanical and chemical dust control methods, such as dust collectors and chemical foams and binders, are well-known and effective ways to minimize dusting when coals are too dry. With wet coals, coal handling problems can severely impact coke plant operations, even stopping coke production when the coal cannot be transported to the ovens. Until recently, mechanical methods such as air lances and vibrators, along with plant operators equipped with sledge hammers, were the only methods available to prevent wet coal from impacting operations. Recent research in the area of chemical coal flow aids has focused on coking coals, including synfuel and lower grade coals now being used in North America due to the high price and limited supply of prime coking coals. It has been discovered that conventional chemical coal flow aids used to improve coal flow properties in the electric power industry are often not effective, or have a negative effect on some synfuels and coking coals. This has prompted the development of several new chemical coal flow aids designed specifically for synfuel and coking coals currently being used in coke plants throughout North America. THEORY Chemical coal flow aids function by one or more of the following mechanisms: Surface tension reduction - Coal flow aids reduce the surface tension of water on the coal outer surface. This reduces interparticle bridging of coal surface water due to hydrogen bonding, and promotes free drainage. Increased capillary wetting - Chemical coal flow aids allow surface moisture to penetrate deeper into the pores and capillaries of coal particles, reducing the amount of surface moisture and resulting in less interparticle bridging. Increased lubricity - Chemical treatments for improving coal flow minimize adhesion not only between coal particles, but also between coal particles and the surfaces of hoppers, chutes, bins and other coal handling equipment where coal particles tend to adhere. Increased hydrophobicity - Chemical coal flow aids minimize interparticle bridging and adhesion due to the adsorption of surface active molecules onto the coal outer surface. In this case, the hydrophilic head of the molecule attaches to mineral and/or oxidized sites on the coal surface. The hydrophobic tail of the molecule is left exposed to render the coal surface more hydrophobic.

It is difficult to predict how an individual treatment chemical will perform on a given substrate. This is because positive and negative effects may occur simultaneously when using a specific treatment chemical. For example, if a treatment reduces surface tension but also decreases capillary wetting, there may be no improvement or even a net decrease in coal flow properties. Due to the increasing use of synfuels and low grade coking coals in North America, research was conducted on a variety of individual surface-active molecules and blends of these molecules. The goal of this testing was to determine the treatments that best improve coal flow properties on a variety of coking coal substrates. LABORATORY TEST PROCEDURE Coal samples were obtained from a various North American coke plants. The samples were screened and reconstituted to ensure uniform particle size distribution. Surface moisture was added and the coal samples were then allowed to equilibrate at ambient temperature. Coal flow measurements were performed using a proprietary technique developed by GE Infrastructure Water & Process Technologies. Test results were compared to conventional powder flow test methods and to actual field data. The procedure has been found to be a better predictive tool for determining coal flow properties in coal handling operations. It is also a fairly simple and portable method that can be used to evaluate coal samples at individual coke plant locations or in the coke plant quality control laboratory. The coal flow test measures the amount of energy required to move a specified amount of coal, with a known amount of surface moisture, through a conical hopper with a fixed orifice and hopper surface area. If all of the coal is not removed from the hopper

AISTech 2004 Proceedings - Volume I

615

within a given period of time, the percentage of coal moved through the hopper is recorded. Given the energy used and the fraction of material moved, a coal flow index is calculated according to Equation 1. Results are then plotted as a function of surface moisture and/or chemical flow aid dosage as shown in Figures 2 and 3. CFI = 1 / (Em + 1) where CFI = coal flow index, and Em = (% of energy used) / (% of material moved). (1)

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Control

0.25 gallon treatment per ton of coal

Coal Flow Index

Chemical A Treatment

Chemical B

Figure 2 Effect of Chemical Treatments on Coal Flow Index

0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 6.3%

Synfuel feed stock Treatment Rate = 0.25 gal/ton

Coal Flow Index

No Treatment Treatment B Treatment A

8.3% % Moisture

10.3%

12.3%

Figure 3 Performance of Chemical Treatments on Synfuel Feed Stock as a Function of Moisture Content

616

AISTech 2004 Proceedings - Volume I

As shown, these data can predict not only which treatment chemistry will be most effective for a given coal, but also what level of surface moisture can be tolerated at a given treatment dosage before the coal flow index deteriorates. By repeating the tests at various chemical dosages, an algorithm can be developed which would allow required treatment rates to be predicted as a function of coal moisture content. This algorithm could then be used to drive an automated chemical feed system. FIELD EVALUATION PROCEDURE There are many interrelated variables and system dynamics which influence the performance of an operating coke plant. For this reason, it is necessary that these variables be understood and considered when evaluating chemical coal flow aids. It is equally important that the specific goals of the treatment program be agreed to and documented prior to any in-plant chemical evaluation, and that any measurement techniques be validated and understood by plant operating personnel. Data must be analyzed carefully according to accepted statistical methods so that treatment program effects may be separated from normal process control variation. Once improvements are made to the process, proper monitoring and control techniques must be developed, documented, and implemented in order to maintain the improved control. One method that has been used successfully to identify and evaluate process control improvements is the Six Sigma approach. This is a formalized method adopted by several major corporations which enables process control improvements to be made via a defined stepwise approach, using statistical techniques in order to quantify process variation and results. While it is beyond the scope of this paper to discuss tedious statistical computations or experimental design, the process used is relatively simple to understand. The approach consists of five basic steps as described below: Define A project team is formed, specific project goals are agreed to, and these goals are documented. Important process variables are identified. Measure Performance standards are defined and the measurement system is developed and/or validated. Analyze Performance objectives are documented, current process capability is evaluated, and potential sources of variation are identified. Improve Sources of variation are screened to determine those with the most influence on the process. Variable relationships are studied and operating tolerances are determined. Control New, improved process control capability is measured and controls are implemented.

The ultimate goal of the Six Sigma approach is to achieve world class control over a particular process. Many processes operate with a level of variation such that there are too many defects, resulting in higher operating costs and lost production. For a business or manufacturing process, the sigma or Z-value of the process measures the ability of the process to perform defect-free work. A defect is defined as anything that results in customer dissatisfaction. In general, as the number of defects decreases, the sigma of the process increases. To illustrate this, Table I shows how a change in going from 99% quality (3.8 sigma) to six sigma would reduce the number of defects in several known processes. Table I Process Defects and Sigma Level 99% Quality or 3.8 Sigma 20,000 per hour 15 minutes per day 5,000 per week 200,000 per year 7 hours per month

Defect Lost Articles of Mail Unsafe Drinking Water Incorrect Surgical Procedures Wrong Prescriptions Power Outages

99.966% Quality or 6 Sigma 7 per hour 1 minute every 7 months 1.7 per week 68 per year 1 hour every 34 years

Both of the case studies presented in this paper utilized the Six Sigma process as a basis for evaluating chemical coal flow aids. The process was found to be very effective in driving the projects forward in a consistent, logical, and quantitative fashion.

AISTech 2004 Proceedings - Volume I

617

CASE STUDIES In this section, two case studies are presented which illustrate the effectiveness of chemical coal flow aids on coke plant production rates. These case studies were performed at two coke plants operated by Stelco. Stelco Hamilton Stelco Hamilton, located in Hamilton, Ontario, operates one coke oven battery consisting of 83 ovens. The No. 7 Battery at Stelco Hamilton was constructed in 1972, and was refurbished in 1998 for major end flue repairs and other system upgrades. The plant produces 2,000 net tons/day of coke and holds the distinction of being the longest coke oven battery in North America. The coke plant at Stelco Hamilton uses a blend of Western Canadian and Appalachian coals as feedstock. Historically there have been some problems, especially during the wet spring and fall months, with coal plugging throughout the coal handling system. This has caused delays in filling ovens, and resulted in reduced coke production rates. A Six Sigma project was undertaken in order to reduce the frequency of coal pluggage in the Stelco Hamilton coal handling system. Using the Six Sigma format and methodology discussed previously, it was agreed that the critical variable to be improved would be delay minutes in filling coke ovens. It was also agreed that for this project, a defect would be defined as any event which resulted in a delay in filling coke ovens. Realizing that a variety of plant operational practices and even plant design limitations could affect coke production delays, these changes were considered to be outside the scope of the study. Since it is known that coal moisture content affects coal handling properties, the moisture content of the coal was measured during the project. Statistical analysis of baseline data showed that the baseline process capability was 0.109, meaning that there were roughly 109,000 defects per million opportunities. In other words, for every million days of operation, we could expect the baseline process to have 109,000 individual delays in filling coke ovens. In statistical terms, this relates to a sigma level of 2.79 which is considered to be below the zone of average process control (3.0 4.5 sigma). It can also be noted that the process appears to be operating well below the accepted zone of world class technology (6 sigma). Figure 4 graphically depicts the current process capability as a function of sigma.

Control Variation
3.0 2.5 2.0

Zone of Average Technology

Zone of

1.5 1.0 0.5 0.0

Typical Control

"Sigma" or Z-value
Figure 4 Stelco Hamilton Baseline Process Capability

618

AISTech 2004 Proceedings - Volume I

In order to improve the process capability at Stelco Hamilton, it was recommended that a proprietary chemical additive be applied to the coal in order to improve coal handling properties. The plant selected a chemical treatment approach as an alternative to system design modifications which would have required significant capital expenditures. The proposed liquid product was diluted with a known ratio of water and mixed with air to create a foam, using an automatic feed device specifically designed for this purpose. The foam was applied to screened component coals as they were being transferred via conveyor to individual mixing bins. The foam application technique was utilized to insure maximum coal coverage while using a minimum amount of chemical product. A diagram of the coal handling system showing the chemical application point is shown in Figure 5.

C1 Underground Hoppers 1-5 C6 Screens

Mixing Bins

Crusher

Chemical Coal Flow Additive C2 C10

#7 Battery Coal Bunker

Figure 5 Stelco Hamilton Coal Handling System and Chemical Application Point Chemical feed was initiated for two weeks between December 16 and December 30, 2003. The trial was stopped in January 2004 due to frozen coal which is known to flow well in the Stelco Hamilton coal handling system. However, a significant number of delays were experienced in January due to coal freezing in ground hoppers, not related to normal coal flow or coal handling properties. The trial was restarted in February 2004 and data was collected through mid-April 2004. Figure 6 summarizes the measured coke oven filling delay frequency during the course of the project along with coal moisture content. The data suggests that treating the coal may significantly reduce the number of delays in filling coke ovens at Stelco Hamilton. When comparing delay frequency during the trial to the same period in 2003, there is a marked reduction in delays during the periods when chemical was being applied. It should also be noted that during the trial, coal moisture content was higher than during the same period in 2003 as shown on Figure 6. Since it is well established that high coal moisture content negatively impacts coal handling properties, it is noteworthy that there were fewer delays with chemical treatment even when higher moisture coals were being used. Statistical process capability was computed for the trial phase of the project. The results show an improved process capability over the baseline case, with a sigma level of 3.2. This relates to roughly 42,000 defects per million days, an improvement of over 60% when compared to the baseline process capability. This process improvement also moves the process into what is considered to be the zone of average process capability as shown on Figure 7. The results of the chemical treatment evaluation and Six Sigma project at Stelco Hamilton are very promising. Further work needs to be done in order to complete the entire project and optimize process operation. For example, the methodology for measuring delay times needs to be evaluated from a statistical standpoint and process controls need to be developed and implemented. In addition, it would be desirable to evaluate the chemical treatment over a longer time period to further quantify results.

AISTech 2004 Proceedings - Volume I

619

Number 40 10.1 9.9 10.0 9.4

% Moisture 12 10.1 10

35 8.5 8.6 8.1 8.6 8.7 7.9 7.9 8.2 8.6 9.0

9.5 8.2

9.7

30

8 25

20

Frequency

15

% Moisture
4

10 2 5

0 Jan- Feb- Mar- Apr- May- Jun- Jul-03 Aug- Sep- Oct- Nov- Dec- Dec- Jan- Feb- Feb- Mar- Apr03 03 03 03 03 03 04 04 03 03 03 03 03 03 04 04 04 Trial Trial Trial Trial Date

Figure 6 Stelco Hamilton Coke Oven Fill Delay Frequency and Coal Moisture, January 2003 to April 2004

Control Variation
3.0 2.5 2.0

Zone of Average Technology

Zone of

1.5 1.0 0.5 0.0 0 1 2 3 4 5 6

Typical Control

"Sigma" or Z-Value
Figure 7 Stelco Hamilton Process Capability Using Coal Flow Aid

620

AISTech 2004 Proceedings - Volume I

Stelco Lake Erie Stelco Lake Erie is located on the north shore of Lake Erie in the village of Nanticoke, Ontario. The No.1 Coke Battery at Stelco Lake Erie was commissioned in 1980 with major upgrades being performed in 1998. The battery consists of 45 ovens of 6.7 meter height, and is the oldest tall battery still standing in North America. Plant capacity is roughly 540,000 metric tons per year. Similar to the Stelco Hamilton coke plant, the Lake Erie facility uses a combination of Western Canadian and Appalachian coals as feedstock to the coke ovens. During cold weather and wet seasons, coal handling becomes a problem due to wet and freezing coal. This directly impacts coke plant production rates. A Six Sigma project was undertaken at Stelco Lake Erie. For this project it was agreed that a defect would be defined as a dropped oven caused by coal flow. Shift foreman reports were used to measure the number of dropped ovens in order to determine process capability. The stated goal of the project was to obtain a positive shift in process capability of 1.5 sigma. As illustrated on Figure 8, statistical analysis of baseline data indicated that the process capability without chemical treatment was at the 4.5 sigma level, which relates to 1,428 dropped ovens per million attempts. As shown on Figure 8, a 4.5 sigma process capability is considered to be on the high end of the range for processes operating with average technology.

Control Variation
3.0 2.5 2.0

Zone of Average Technology

Zone of

1.5 1.0 0.5 0.0 0 1 2 3 4 5 6

Typical Control

"Sigma" or Z-value
Figure 8 Stelco Lake Erie Baseline Process Capability In order to improve the process capability, it was suggested that a chemical coal flow aid be evaluated. Based on the coal handling system design, it was decided that the best chemical application point would be to the C6 conveyer just ahead of the crusher. Treatment chemical was mixed with air and applied as a foam to obtain maximum coal coverage. Figure 9 provides a schematic of the Stelco Lake Erie coal handling system, and indicates the location of the chemical feedpoint. Statistical evaluation of treatment results during the chemical trial period revealed a significant improvement in process capability. Figure 10 shows that the new process capability with chemical treatment improved by 1.5 sigma (the defined goal of the trial), resulting in a process that achieved greater than 6 sigma performance. Using a sample size of 2,614 oven fill attempts during the chemical evaluation period, there were no dropped ovens as a result of coal handling problems. During the trial period, there was essentially no requirement for manual intervention in order to clear coal blockages in chutes or hoppers throughout the coal handling system. Previous to the trial, and without chemical treatment, manual methods were often employed in order to clear coal blockages.

AISTech 2004 Proceedings - Volume I

621

Chemical treatment results at Stelco Lake Erie exceeded expectations with regard to coal flow improvement. In order to complete the Six Sigma project, a process control plan still needs to be implemented. It is also planned to automate the chemical feed system so that treatment chemical usage and system performance can be optimized.

C4

C3 C5

C2

C1

C6

Chemical Coal Flow Additive

Screens

South Crusher

C7 C8 C9

C10
5 3

Oil
4 2 1

Silos

C11 Mixer C12 Chute Bunker C13 C13 Chute Bunker

Figure 9 Stelco Lake Erie Coal Handling System

Control Variation
3.0 2.5 2.0

Zone of Average Technology

Zone of

1.5 1.0 0.5 0.0 0 1 2 3 4 5 6

Typical Control

"Sigma" or Z-value
Figure 10 Stelco Lake Erie Process Improvement Using Chemical Coal Flow Aid

622

AISTech 2004 Proceedings - Volume I

SUMMARY The case studies presented in this paper indicate that chemical coal flow aids can be highly effective in reducing delay times or dropped ovens in the manufacture of coke. As of this writing, both of the Stelco locations studied in this paper continue to use the coal flow aid chemical treatment technology and both locations are operating at record coke production rates. Given the current high value of coke and the continuing upward price pressure, maintaining high production rates results in increased coke plant profitability. While it is acknowledged that many variables affect the coking process, it is obvious that if coal is not getting to the ovens, production rates will suffer. By using the Six Sigma approach to analyze the coal flow problems at Stelco, the results of introducing the chemical treatment technology could be directly quantified. Without using this approach, it would have been difficult, if not impossible, to isolate the process variables and measure the effects of coal flow treatment on the process. While the case studies presented here did evaluate the effect of the chemical treatment on production rates, there are other potential benefits of the treatment program that will require further study. These include possible reductions in the requirement for bulk density oil, minimization of damage to coke ovens due to improper filling, improvement in coke quality, and possible downstream benefits on the coke by-products plant. REFERENCES 1. Energy Information Administration, Annual Coal Report 2002, U.S. Department of Energy, Washington, D.C., 2003. 2. Xinhua News Agency, Government Warns of Overheating in the Coke Industry, Press Release, May 24, 2004.

AISTech 2004 Proceedings - Volume I

623