英文原文—劉小川2007.pdf

OPERATIONAL EXPERIENCE FROM THE UNITED STATES FIRST VERTICAL ROLLER MILL FOR CEMENT GRINDING By: Mark Simmons Production Manager Salt River Materials Group Phoenix Cement Plant Lee Gorby Manager of Quality Control Salt River Materials Group Phoenix Cement Plant John Terembula Product Manager Milling FLSmidth, Inc. ABSTRACT For several decades the cement industry has successfully utilized Vertical Roller Mills (VRM) for grinding of raw materials and solid fuels. Most recently, this technology has been employed for the comminution of Portland cement, blended cements and slag cements. The VRM offers several benefits compared to the ball mill in regards to operating costs and flexibility. However, the quality of the cement produced is extremely important in cement grinding and there is little experience with cement produced from a VRM in the US market. This paper relates the operational experiences from the first VRM for clinker grinding put into operation in the United States in 2002. Included in the discussion are operational data, maintenance discussion and laboratory data focused on product quality. All of the discussion is based on comparison to ball mill operation at the same plant. INTRODUCTION Traditionally, the closed circuit ball mill with high efficiency separator has been the most common system for cement grinding. However, as happened with raw grinding over the last 25 years, the vertical roller mill (VRM) is now successfully being used for many clinker grinding applications and is rapidly becoming the standard for new grinding installations. The first such vertical roller mill installation in the United States was part of a total plant expansion and began operation in August of 2002. Phoenix Cement Company modernized their existing three kiln plant to a state-of-the-art high efficiency operation by incorporating the newest technology available in the cement industry. Spurred by a growing demand for cement, the cement producer increased their clinker capacity from 1700 STPD (1590 MTPD) to a capacity of 3000 STPD (2700 MTPD). For the new VRM cement grinding system, the cement producer contracted with an equipment supplier based in Bethlehem, Pennsylvania for a proven mill design that was originally developed in Japan during the early 1980s. FIGURE 1 shows the VRM installed in the plant. FIGURE 1: Vertical Cement Mill Near Phoenix, AZ 0-7803-9107-1/05/$20.00 (c)2005 IEEE CEMENT GRINDING IN A VERTICAL ROLLER MILL The differences between raw and cement grinding have been well documented in numerous publications and presentations over the recent past. Specifically, as compared to limestone, clinker and cement raw materials are finer and harder to grind. This, coupled with the finer and more stringent product particle size distribution requirements, entails design considerations to allow for continuous and stable operation of the grinding system. The patented geometry of the mills grinding parts has demonstrated its suitability for this application at the Arizona cement plant. As shown in FIGURE 2, the rollers are spherical in shape with a groove in the middle. The table is also curved forming a wedge-shaped compression and grinding zone between the rollers and the table. FIGURE 2: VRM Patented Mill Table and Roller Profile The dual-lobed design is optimal for clinker grinding because it supplies two distinct grinding zones, a low pressure zone and a high pressure zone, at each roller. The low pressure area under the inner lobe de-aerates and consolidates the material to be ground. This ensures a compact well established grinding bed for maximum stability. The proper grinding then takes place in the high pressure zone under the outer lobe. The groove in the middle of the roller facilitates de-aeration of the material without fluidizing it. Due to its design of grinding parts and integral high efficiency separator, the vertical cement mill, as shown in FIGURE 3, addresses all the difficult grinding conditions associated with the fine grinding of cement clinker and related products. The result is a high grinding efficiency and extremely stable mill operation. 0-7803-9107-1/05/$20.00 (c)2005 IEEE Separator Grinding roller Rocker arm Gear box Mill stand Separator drive Mill outlet Mill feed inlet Grinding table Hydraulic system FIGURE 3: VRM for Cement PROCESS The cement producer, headquartered in Phoenix, is a regional supplier of Portland cements, gypsum, and fly ash products. Their vertical roller cement mill is one of the most modern components of the plant. Although the new mill was rated for 130 STPH (118 MTPH), it has operated consistently in the range of 140-150 STPH (127 to 136 MTPH) at a product fineness of 3900-4000 Blaine. FIGURE 4 shows the cement mill production both immediately after commissioning and after 2 years of operation. 100.0 120.0 140.0 160.0 180.0 Aug-02Dec-02Apr-03Aug-04Dec-04 Production (STPH) FIGURE 4: Cement VRM Production 0-7803-9107-1/05/$20.00 (c)2005 IEEE Maximum capacity of 165 STPH (150 MTPH) was achieved shortly after commissioning. General operating conditions do not require this capacity on an ongoing basis; however the availability to meet a future increase in demand has been proven. In addition to increased capacity, a significant power savings was realized on the order of 16 kWh/MT of finish cement. As seen from the data in TABLE 1, the VRM for cement grinding offers a significant advantage in power savings. Typically, the VRM uses 50% less power than a ball mill when grinding the same clinker to 3900 Blaine. VRMs are also much more adept at handling hot feed compared to ball mills. The simple and compact vertical mill layout is cost competitive to build and offers many options for layout, even in existing plants. Today significant operating experience has been accumulated with vertical mills ranging from plant design and layout to operation with multiple types of product. One of the main focus points regarding cement VRM operation in the USA has been product quality and the product compatibility with existing ball mill systems. QUALITY In general the cement grinding system is required to consistently grind the prerequisite capacity of cement product while meeting all established product quality standards. The numerous quality parameters measured in a cement plant are assigned varying levels of importance depending on geographic region, individual plant operators and the needs of specific customers. Further, it is safe to say that consistency is equally important as obtaining any quality parameter. The most commonly used measure of the mill system is product fineness. Based on established conditions under which the product yields the desired reactivity and consequently strength development, a standard particle size can be used to monitor mill consistency on an hourly basis. This usually takes the form of a single sieve residue or more often the specific surface (Blaine). Because the effects of the mill on the quality of the product are more complex than what is practically reflected by a Blaine value or a sieve residue, the following product quality factors should be measured and monitored on a regular basis: The particle size distribution (PSD) and Blaine The degree of dehydration of the gypsum added to the cement The prehydration and carbonation of the clinker minerals Various feed materials are known to behave differently from each other as a result of inherent differences in material characteristics. However, materials will also react differently according to the many types of grinding systems in which they are processed. A considerable data base of Ordinary Portland Cement: 95% clinker, 5% gypsum VRM Ball Mills Production tph 144 31 Mill Diff. Pressure mm WG 494 - Mill vibration mm/s 1.3 - Grinding aid % 0.018 0.018 Blaine target cm2/g 3900 3880 Specific energy consumption: Mill kWh/t 18.3 35.2 Classifier KWh/t 0.36 4.7 Fan kWh/t 7.1 2.5 Total kWh/t 25.7 42.4 TABLE 1: Cement Mill Operating Data 0-7803-9107-1/05/$20.00 (c)2005 IEEE ball mill operation has been established throughout the history of ball mill cement grinding. The deficiencies as well as adequacies of ball mills are well known with regard to quality issues making the effect of each quality issue critical to understanding the VRMs reliability as a cement mill. QUALITY CONTROL In a cement vertical roller mill grinding is performed in closed circuit and with an integral high efficiency separator. This arrangement will give a good steep PSD. Experience has shown that the overall product particle size distribution is consistent with that obtained from a ball mill grinding plant with a modern high efficiency separator. During the initial VRM optimization period the mill is fine-tuned to match its product to the existing ball mills. This is achieved by making adjustments to operational parameters such as: Separator rotor speed Air flow rate Grinding pressure Dam ring height Because the VRM has significantly higher grinding efficiency than a ball mill there is much less heat input from the grinding process. This is evident in the almost 50% less installed power, but is further taken into account with a smaller percentage of the energy being absorbed by the material. Compared to ball mills where 75% of installed power may be absorbed a good VRM will take only 50% of the installed motor power as heat. The end result is that the product will not be heated up as much as in a ball mill. This means that a lower degree of gypsum dehydration could occur. A lesser degree of gypsum dehydration is not problematic considering two conditions; the inability to adequately control temperatures in ball mills creates an environment where operation is at the extreme of the gypsum dehydration. Additionally, less dehydration is not an issue if the gypsum is sufficiently reactive to control the setting reactions with a lower degree of dehydration as is normally the case. If in special cases this is not the case different options are available to cope with the problem: Addition of more gypsum (within the SO3 limit) Increased dehydration of gypsum by adding more heat to the mill system Addition of a more reactive form of gypsum Prehydration is not typically problematic in a VRM as it is in ball mill systems where higher temperatures and internal water-cooling systems are common. However, if cement is produced at a relatively high temperature and still has a lot of gypsum that is not dehydrated one must be aware of the potential problem of gypsum dehydration coupled with clinker prehydration that can take place during storage in the cement silos. If a problem of this kind is present it can be coped with it by one (or more) of the following options: Ensuring that the cement is cooled to a lower temperature before going into the silo Provoking a higher gypsum dehydration level in the mill Replacing part of the gypsum with natural anhydrite The actual VRM results achieved in Phoenix are built upon the theoretical rules for quality control presented here. During the initial period of operation, extensive quality data was recorded and analyzed. The results from comparisons between the existing ball mills at the Phoenix plant and the new VRM are presented in an abbreviated form. The data summarized below are based on 0-7803-9107-1/05/$20.00 (c)2005 IEEE significant results to comply with all plant requirements, but more importantly to conclusively satisfy all market requirements. VRM AND BALL MILL PRODUCT COMPARISONS To compare the similarities or differences between cement mill products of the same composition but produced in a vertical roller mill or ball mills, samples from each of the plants mill types were taken and tested. All samples were obtained while the mills were producing the same product. The existing mills are three circa 1950s ball mills from 2 different original suppliers. Two of the ball mills are 12 feet in diameter and 18 feet in length and each mill has twin 14 foot separators. These mills have 1500 horsepower motors and are rated at 30STPH. The third ball mill is a double compartment mill measuring 9 feet in diameter by 33 feet long with a single 16 foot dynamic separator. This mill also has a 1500 horsepower motor and is rated at 30STPH. Particle Size Distribution TABLE 2 presents a comparison of product particle size distribution from the 4 mills. Due to the inherent flexibility of the VRM and separator notice that the product fineness for the VRM does not match exactly, however key target residues are matched. More details of the balance between matching PSD and strength development follow. 100um 45um 30um 10um 1um 0.6um Blaine Ball 1 1.0 96.6 73.7 39.0 2.5 0.78 3900-4000 Ball 2 1.0 96.7 74.4 39.2 2.9 0.65 3900-4000 Ball 3 1.2 94.0 70.1 38.5 2.8 2.3 3900-4000 VRM 0 95.8 68.8 40.7 4.7 2.2 3900-4000 TABLE 2: Product Particle Size Distribution Data Strength Development Strength development data for product samples from all 4 mills were compared. FIGURE 5 shows comparative ASTM C109 cube strength results for the ball mills versus the vertical mill at two different Blaine targets. Notice the VRM has generally better strength development. 0-7803-9107-1/05/$20.00 (c)2005 IEEE 0 1000 2000 3000 4000 5000 6000 7000 Compressive strength ASTM C109 (psi) 1 Day3 Day7 Day28 Day BM VRM FIGURE 5: Compressive Strength Data As mentioned above in the Quality Control section the gypsum dehydration and PSD will contribute to the cement strength development. By having the flexibility to adjust the operating parameters such as separator speed, temperature and dam ring height the PSD and dehydration can be controlled independently to achieve the necessary balance for strength development. The PSD from the vertical mill may not exactly match that of the ball mill; however the specific strength development will either match or show improvement. The improvement obtained in strength development was not possible with the ball mills due to inherent operational limitations. 0-7803-9107-1/05/$20.00 (c)2005 IEEE CONCLUSION Two years of cement vertical roller mill operation have proven the decision to invest in this new grinding