INCREASING TOP ROLL SHAFTS RELIABILITY. J.J. Coronado*, S.A. Rodriguez*, A.L. Gómez*, J.S. Rivas*, C.A. Velez**

Transcription

1 INCREASING TOP ROLL SHAFTS RELIABILITY By J.J. Coronado*, S.A. Rodriguez*, A.L. Gómez*, J.S. Rivas*, C.A. Velez** *Mechanical Engineering School Universidad del Valle - Colombia,**Manuelita S.A.Mill. johncoro@petecuy.univalle.edu.co, sararopu@hotmail.com, fito@petecuy.univalle.edu.co, jsrivas@petecuy.univalle.edu.co, cvelez@manuelita.com KEYWORDS: mill shaft, fracture, welding. ABSTRACT Several top roll shafts failed in service in different Sugar Mill and after analysis it was determined that failures happened due to a fatigue process, where the cracks started at surface located microdefects, originated by abrasive wear, machining marks or defects of the repair welds when applied. A research project to extend the life of top roll shafts presenting excessive wear and developing cracks was structured with Mechanical Engineering School of Universidad del Valle (Cali Colombia), CENICAÑA and Manuelita Mill. Failures location and frequency were documented together with materials, torque and crushing loads. Shaft journal on the drive side was chosen for detailed analysis because the higher frequency of failures presented there. Two options were considered to obtain a shaft free of cracks: the reduction of the journal diameter and the recovery of the original dimensions using arc welding procedures and consumables. A Fracture Mechanics approach was used to predict the maximum defect size tolerated in a repaired shaft and time between Non Destructive Testing (NDT) inspections. PALABRAS CLAVE: eje de molino, fractura. soldadura. RESUMEN Los ejes superiores de molino de caña de azucar presentan fallas causados por un estado de fatiga, donde las grietas surgen en la superficie del eje, debido a defectos originados por desgaste abrasivo, marcas de maquinado o defectos originados por recuperación con soldadura. Un proyecto de investigación para aumentar la vida de ejes de maza superior que presentan excesivo desgaste y grietas fue desarrollado por la Escuela de Ingeniería Mecánica de la Universidad del Valle, Cenicaña y el Ingenio Manuelita. La frecuencia de las fallas fue documentada junto con materiales, torques y cargas. El guijo del eje de maza superior fue seleccionado para un análisis detallado debido a la alta frecuencia de fallas que presenta dicha zona. Se consideraron dos métodos para obtener un eje libre de grietas: reducción del diámetro del guijo y recuperación de las dimensiones originales usando diferentes depósitos y procesos de soldadura. Una aproximación de la Mecánica de Fractura fue usada para predecir el tamaño de grieta máximo permitido y el tiempo entre inspecciones en ejes recuperados por soldadura usando pruebas no destructivas. INTRODUCTION A cane crushing mill is a heavily loaded piece of machinery and because of function and design, top roll shafts are highly stressed elements affected by juice and soil contamination. Several top roll shafts failed in service at Colombian sugar mills in a relatively short period of time and after analysis it was determined that failures happened due to a fatigue process, where the cracks arise from surface located microdefects, originated by abrasive wear, machining marks or defects of the repair welds when applied. A research project to extend the life of top roll shafts presenting excessive wear and developing cracks, dealing with the maintenance cost reduction, was structured

2 by Research Group for Industrial Improvement of the Mechanical Engineering School of Universidad del Valle (Cali Colombia) together with CENICAÑA and Manuelita Mill. From crack detection reports, in 1045 top roll shafts of a 84 x 43 tandem at Manuelita Sugar Mill, both frequency and location of cracks appearance were found, Fig. 1 shows results: 82 failed shafts were reported where 115 cracks were NDT detected. The highest percentage of defects (35 %) was located at journal shoulder close to roll shell, drive side. Different researchers, Anderson & Loughran (1999), Reid (1989), have found similar results. Due to this, loads and stresses in the mentioned zone were considered in this study. Percentage of failure 40% 35% 30% 25% 20% 15% 10% 5% 0% 26% 1% 15% 35% 21% 6% 17% 18% 1% 0% 0% Square end Crown Journal Shoulder Shell - left Shell - middle Shell - right Shoulder Journal - left Journal - right Crown Shaft zones Fig 1. Roll Shaft Cracks distribution at Manuelita S.A TYPES OF CRACKS IN THE MILL ROLL SHAFTS Two crack models agree with experimental studies, Arzola (2003): the semi-elliptical superficial cracks, and the circumferential crack. Several semi-elliptical cracks can arise, in a given section of the shaft, at different times, and they can propagate until become a circumferential crack when they coalesce, Fig. 2 shows a fractured shaft section where different semi-elliptical cracks had propagated until they formed a circumferential crack. Besides it, a circumferential crack can arise and propagate due to circumferential surface defects on shaft journal, for example, deep groove marks cause by machining process or abrasive wear, Fig. 3 shows some marks on the shaft surface journal.

3 Fig. 2. Several semi-elliptical cracks Fig. 3 Abrasive wear on top shaft surface Fracture Toughness of steel used on roll shaft. Standard CTOD tests were performed on 1045 steel in order to determine fracture toughness. A device adapted to a universal milling machine was used for specimens precracking. A figure of 63 Mpa-m 1/2 was determined, Hinestrosa O. (2004) Rotating bending fatigue and juice corrosion synergy for 1045 steel. Rotating bending fatigue and juice corrosion synergy for 1045 steel was considered, Gómez J.A. (2004) using a pre- immersion, loading, immersion and final loading of standard rotating fatigue specimens. Typical fourth mill degraded juice with a 4 PH was used for trials. During rotation a continuous flow of juice was also applied. Comparison between dry standard test and the adapted corrosion experiment showed a reduction of 15% in fatigue limit, Sn. This results could be applied for zones not protected by lubricant presence. In the analysis of journal shoulder a reduction of fatigue life was not considered. Typical arc welded metallography analysis A welding recovered shaft by was analyzed to study the fracture cause. Welding deposit with austeno-ferrític structure is showing in Figure 4, where the dendrites formation is observed clearly. In the Figure 5, a crack was observed in the interface between the base metal and the weld metal, coming up from a superficial defect supposedly by a wrong procedure. The base metal preparation, shielded electrode diameter selection as well as the welding parameters were taken into account. A lack of fusion was observed clearly In Figure 6, allowing a selective crack development direction. The interface between ferritic- perlitic base and austeno-ferritic weld metal is presented in Figure 7.

4 Figure 4. Recovered welding. Figure 5. Cracks in recovered welding. Figure 6. Welding defect. Figure 7. Base metal and welding interface. Loads on a roll shaft Journal loads and stresses vary with lift and milling torque. Hydraulic lift is an approximately constant load system but torque sharing inside the mill is a complex process. Torque load recordings for a shaft with a 19 x 26 journal were taken, Rodriguez S. (2004) to establish fractions of time corresponding to load levels. Shaft material was A history of radial force generated in squared end of a top roll shaft because of bending moment induced by square coupling was calculated from measured torque using work of Ókamura H. et al, (1972), Figure 8.

5 Figure 13. Estimated radial force in shaft squared end from real torque measurements.. There were selected 5 radial representative ranges of force and the corresponding information was grouped in every range. Table 1 shows ranges and corresponding fractions of time as well as the torque and lift information. General stress state of a top roll shaft was then calculated using all internal loads and a typical load share among rolls. Table 1. Radial Forces on squared end CONSIDERATIONS TO REPAIR A SHAFT JOURNAL WITH EXCESSIVE WEAR OR CRACKS IN PROGRESS Plastic zone size in the crack tip. Plastic zone in the front of crack must be eliminated in order to stop crack propagation. This plastic zone will always have a size that can be estimated according to the stress intensity factor (K) and the yield stress (σ y ). The size of the plastic zone under plane strain condition can be obtained as: r y 1 K = 6 π σ y 2 Figures 9 and 10 show plastic zone size for a semielliptical and circumferential cracks respectively.

6 Figures 9. Extension of the plastic zone for a semielliptical superficial crack in the journal shaft. Figures 10. Extension of the plastic zone for a circumferencial crack in the journal shaft. Using fracture toughness, Kic in the equation a value of 3.35 mm was obtained as maximum extension for the plastic zone in plane strain condition in the 1045 steel. Therefore in a recovery process it is required to eliminate this zone in the crack tip. Journal diameter reduction When a journal shaft presents heavy wear caused by the high stress and the contamination with extraneous matter, machining to give a new diameter is an alternative for the shaft to continue working. One of the questions that must be resolved is what is the minor diameter to which it is possible to take a top shaft before having to recover the original size using arc welding. The new value for journal diameter raises nominal stresses as can be observed in Figure 11, besides, it increases the effect of the stress concentration caused by the change of section between the journal and the zone of shell landing as can be observed in Figure 12 where bending and torsion stress concentration factors increase both with increased diameters relation. Figure 11. Equivalent stresses in the journal for different diameters.

7 Journal Diameter (m) 0,48 0, Journal Diameter (m) 0,48 0, Kf 1.13 Ks Bending Shaft Depth (m) Shaft Depth (m) Torsion Figure 12. Stress concentration factors Kf and Ks for different journal diameters. The critical crack size on semielliptical and circumferential cracks for different values of diameter can be seen in the Figure 13. Semieliptical crack Circumferential crack Figure 13. Critical crack size for different diameters The critical crack size cannot be used as a criterion for the determination of minimum feasible journal diameter. It will depend on the time that takes to a crack with the minimum size to be detectable, to propagate up to the critical length. This value can be established using as a criteria the period between NDT inspections. Figure 14 shows a month as the time to inspect the evaluated top roll shaft with reduced diameter. This result is unpractical so a more feasible time between NDT examinations will define the lowest diameter that must be used.

8 Figure 14. Minimum time between NDT inspection to define the shaft diameters. Shaft Diameter restoration by arc welding. Different arc welding procedures have been used to repair roll shafts. These procedures also have been designed for typical 1045 steel shaft. One of the procedures used in Manuelita S.A. sugar mill consists uses submerged arc welding (SAW) with neutral flux. The main advantages of SAW consist in high deposition rates, and the low hydrogen embrittlement sensibility, if the flux is correctly handled. A new stress distribution can be calculated for shaft surface by using a 13 MPa residual stress value for 35.2 Kg/mm 2 welding, after 4 hours of heat treatment and superposing its effect to the operational stress, Figure 15. Load state 1 Load state 5 Figure 15. Normal and shear stress distribution in the journal after a stress relieving treatment.. The same repaired shaft without stress relieving was considered. Residual tension welding stresses of 170 Mpa for 35.2 Kg/mm 2 welds were superimposed to operational stresses on the shaft. Residual stresses acting all around the repaired zone were assumed. Stress distribution can be observed in Figure 16. The calculated critical crack size was 9 mm for semielliptical and 8 mm for circumferential cracks. A time between NDT inspections of 40 days was also estimated. Calculated inspection periods can be extended a 18% if minimum detectable crack size (depth) of 3 mm is considered instead of 4 mm used in this calculations. Currently CTOD measurements are being performed on specimens taken from a real shaft with 16 x 21 journal where four different consumables were applied with two SAW and two MIG procedures.

9 Stress (Pa) 4,E+08 3,E+08 3,E+08 2,E+08 2,E+08 1,E+08 5,E+07 Normal Shear Load state 1 Load state 5 Figure 16. Normal and shear stress distribution without stress relieving. 0,E Degrees CONCLUSIONS Maximum stresses are located at journal shoulder close to roll shell, drive side. 35 percent of cracks were detected in this location. A direct relationship between lift and shaft journal stresses was found. The critical condition is when roll shaft is above low speed gear shaft. A plastic zone of 3.35 mm in front of crack was estimated for studied shafts. It must be removed before welding. Minimum diameter for a reduced size shaft was calculated. A 5% reduction reduces inspection period a 38% Reliability of a welded shaft can be impaired if a sound stress relieving process is not used. Service life can be reduced in an estimated 63% because residual stresses. Careless welding procedures are fatal for repaired shafts. This repairs, if adopted, need to take into account all the technical considerations and to follow all the required welding protocols. References Reid, M. J. (1989). Possible causes of recent roll shaft failures in South African sugar mill, Proceeding of the XX ISSCT Congress, Sao Paulo. Brazil. Anderson S.I & J.G. Loughran. (1999) Mill roller design and operational stress states, Proc. Aust. Soc. Sugar Cane Technol. Rodriguez P., S.A. (2004), Predicción de vida remanente en ejes de maza superior de molinos de caña. Proyecto de Grado Universidad del Valle. Escuela de Ingeniería Mecánica p Hinestrosa O. (2004) Diseño y construcción de un banco prefisurador de probetas de flexión en tres puntos para determinar la tenacidad CTOD del acero Escuela de Ingeniería Mecánica Universidad del Valle. p. 91 Gómez J.A. (2004), Análisis del comportamiento a fatiga de un acero para ejes en ambiente corrosivo. Proyecto de Grado, Universidad del Valle. Escuela de Ingeniería Mecánica p. 55

10 Ókamura H. et al, (1972). Square box couplings in cane mill drives, International Sugar Journal. Arzola, N. (2003). Esquema de análisis para los árboles de los molinos de caña de azúcar y aplicación de la Mecánica de la Fractura en la evaluación de la falla por fatiga, Tesis PhD Universidad de Cienfuegos, Cuba.