Transactions on Modelling and Simulation vol 15, 1997 WIT Press, ISSN X

Transcription

1 Frit salvage improvement of cathode ray tube by thermal method Chin Young Cha & Shridhar V. Iyer Philips Display Components, Inc., Ann Arbor, Michigan, USA Joseph J. Rencis Mechanical Engineering Department, Worcester Polytechnic Institute, 100 Institute Road, Worcester, Massachusetts USA This paper describes how a thermal stress boundary element analysis was carried out to select a mechanical device that improves the frit salvage of a cathode ray tube (CRT). The panel and funnel of a CRT are fritted at the seal edge. To reuse a CRT, the panel and funnel are separated by etching and then applying a thermal shock. This procedure is known as frit salvage. Current yield of the frit salvage success rate is about 67% which means that 33% of the components are lost and cannot be reused. The financial burden on the company can be reduced if the salvage rate can be improved through having a better understanding of the thermal shock mechanism. During thermal shock, a crack has been observed to initiate at the end of the axes and traverses towards the corner along the seal end. A CRT is lost when the crack travels along the diagonal corner. A thermal stress analysis was carried out using the boundary element method. The analysis determined that maximum stresses are located near the blend radius on the panel skirt at approximately 1-2" on either side of the diagonal. Also it was found that the stresses along the frit seal edge are uniform from the end of the major and minor axes towards the diagonal. The stresses at the diagonal location were reduced by approximately 29%. The direction and location of the crack obtained by the boundary element analysis was consistent with those observed in the frit salvage procedure. A panel corner pullout protector is used at each diagonal corner in order to prevent thermal failure. As a result the maximum stresses at the diagonal corners are decreased significantly. Resultantly, the stresses along the frit seal at diagonal are relatively higher than those on the diagonal corner. Consequently the crack continues to traverse along the high stress lines at the seal edge as required for a successful frit salvage process. This application is an excellent example of using the boundary element method in an industrial setting. 1. INTRODUCTION A CRT is constructed as a continuous envelope of glass. The main features of a CRT are shown by the boundary element model in Figure 1. Numerical techniques for the static stress analysis of engineering components have developed rapidly over the past decade, and today the two major alternatives are the finite element method (FEM) [1]

2 462 Boundary Element Technology XII and the boundary element method (BEM) [2, 3]. Both techniques have been used to carry out a stress analysis of a 27 inch CRT [4, 5]. The boundary element results achieve a high degree of accuracy even for a single element in the thickness direction. Although it is found to be unnecessary to use a more refined mesh, BEM also allows one to re mesh the surface easily in an extremely short period of time. In contrast, the finite element results appear to require a minimum of two eight-node brick elements in the thickness direction before the results approach convergence. The difficulty of re mesh ing a FEM volume mesh is an obstacle in carrying out a rapid convergence analysis. The simplicity of carrying out a refined boundary discretization allows a more detailed convergence study and therefore promotes confidence in the stresses. In consequence, the boundary element model was easier and quicker to build, more accurate and also has an extremely shorter modeling time than the finite element model. 2. PROBLEM The panel and funnel of a CRT are fritted at seal edge. To reuse a CRT, the panel and funnel are first separated by etching the frit along the seal edge by using Nitric acid at 130 F. The glass is then cooled from 130 F to 60 F with water at 60 F in 50 seconds. During thermal shock, a crack has been observed to initiate at the end of the axes and traverse toward the corner along the seal end. However, 33% of the CRTs are lost due to the crack traveling along the diagonal corner as shown in Figure 2a. In order to analyze the mechanism which caused the crack, BEM was chosen due to its ease of use and its well known suitability for solving problems involving cracks and stress concentrations [6]. A thermal stress analysis was carried out using the commercial boundary element code BEASY [7]. Two planes of symmetry are assumed, allowing only one quarter of the full tube to be modeled. A heat transfer analysis was first carried out to determine the temperature variation throughout the CRT during thermal shock. A cooling rate of 84 F /min was assumed to simulate the actual process of cooling from 130 F to 60 F in 50 seconds. A thermal stress analysis was then carried out using the temperature solutions from the heat transfer analysis. All stresses in this paper indicate maximum principal tensile stress. The thermoelastic response of the CRT was analyzed using BEASY. BEASY unfortunately has a limitation which precludes the prediction of stresses generated from distributed volumetric internal heat sources. For this study, an approximate solution was obtained by first determining the surface temperature distribution from a heat transfer analysis which included the volume source, and then applying that temperature portion of the solution for a thermal stress analysis. The assumption was made that the through thickness effects of the volume source is small, and that the term in the elasticity equations which considers the volumetric effects of the source temperature profile is negligible. This assumption is appropriate since the CRT has a comparatively thin section. A quarter symmetry three-dimensional boundary element model of the CRT shown in Figure 1 consisted of 542 continuous quadratic elements and has 8374 nodes. The CRT is assumed homogeneous and isotropic with the following glass material constants (Corning Glass code 9061):

3 Boundary Element Technology XII 463 Young's modulus = E + 6 psi Poisson's ratio Thermal conductivity = = E-5 BTU/in-s F Thermal expansion coefficient = 5.5 E-6 in/in F Reference temperature Volume source = = 68 F E-3 BTU/in* - sec Heat transfer coefficient of water with the ambient temperature 60 F at outer glass surface. = E-3 BTU/in*-sec - F Heat transfer coefficient of air with ambient temperature 130 F at inner glass surface. = 2.55 E-5 BTU/in* sec - F Heat transfer coefficient of air with ambient temperature 60 F at outer glass surface. = 2.55 E-5 BTU/in* sec - F Two different heat transfer coefficients were prescribed on the outer glass surface of the CRT as shown in Figure 1 when the panel pullout protector is used. On the hatched outer surface area the heat transfer coefficient of air was prescribed while the unhatched area water was used. When the panel pullout protector is not used, then the heat transfer coefficient of the outer surface is water. 3. RESULTS The temperature solution over the outer glass surface is shown in Figure 3. The entire outer surface is subjected to a temperature of 60 F. The temperature variation over the inner glass surface is shown in Figure 4. The temperature of the inner face panel is 119 F and gradually decreases to 60 F at the neck area. This temperature variation is very reasonable since the thickness of the face panel is much larger than that of the neck. The principal stress solutions of the outer and inner glass surfaces are tensile and compressjve, respectively, as shown in Figures 5 and 6, respectively. The crack will always begin on the outer glass surface since the stress is tensile and glass is a very brittle material. Figure 7 shows an enlarged view of the stresses over the outer panel surface since this is the crack growth region. A detailed description of the outer panel shown in Figure 2a indicates various locations (A-E) for studying crack growth. The panel's stress analysis results at various outer surface locations for the salvage tube with and without the panel corner pullout protectors are shown in Table 1. The + and - signs in Table 1 indicate the amount of increased and decreased principal stress, respectively, on salvage tube with the panel corner pullout protectors. The principal stress directions on the outer glass surface are shown in Figure 8. The crack starts at point A (on major axis), which has the maximum stress value (3323 psi), and proceeds along the seal edge. The crack then changes direction towards point D instead of point B (seal edge) since the stress of point D is 524 psi greater than point B. This is consistent with Figure 8 since the crack propagates perpendicular to the principal stress direction at point D. The crack then continues to propagate towards

4 464 Boundary Element Technology XII point E and then proceeds to the seal edge and then along the seal edge to point C. This result closely matches the frit salvage process at Phillips Display. Rubber panel corner pullout protector has been attached to the outer heel radius's diagonal corner of the panel as shown in Figure 2b. A thermal shock is applied by pouring water at 60 F on the outer center face of the panel. Water flows over the entire outer surface of the CRT except for the hatched area of the panel and funnel as shown in Figure 1. Water does not flow on the hatched surface area since it is diverted by the corner panel pullout protector as shown in Figure 2b. The panel and funnel has been included as an air cooling portion while air cooling on the yoke and neck surface area below the corner panel pullout protector has been neglected. This assumption is reasonable since the effects of this surface area on the seal edge is negligible due to the large distance of the yoke and neck from the seal edge. The temperature solution of outer glass surface with the panel corner pullout protector is shown in Figure 9. The temperatures of the panel and funnel below the panel corner pullout protector are approximately 107 F with the remaining portion at 60 F. The temperature solution of inner glass surface with the panel corner pullout protector is shown in Figure 10. The temperatures of the panel and funnel below the panel corner pullout protector are relatively higher than the rest of it. The principal stresses of outer glass surface with the panel corner pullout protector are tensile as shown in Figure 11. The stresses of inner glass surface with the panel corner pullout protector are compressive as shown in Figure 12. The principal stress distribution over the outer panel surface with the panel corner pullout protector is shown in Figure 13. These principal stress results yield two significant findings. First, the stress of point D has been reduced to psi, and the stress of point E has been reduced to psi as stated in Table 1. The amount of stress reduction due to the panel corner pullout protector is significant. The principal stress directions at points D and E are approximately 135 and 45 respectively, with the inclination with respect to the seal edge as shown in Figure 14. These directions are consistent to the case without the pullout protector, however, the tremendous reduction of these stresses are against the crack's propagation to points D and E. Secondly, the stress of point A has increased significantly by 1660 psi due to the panel corner pullout protector. This increase of stress promotes crack initiation at point A. Due to the above two findings, the crack will start at point A, which has the maximum stress, proceed to point B on seal edge, which has psi higher than that of point D, and end at point C along the seal edge. 4. CONCLUSION The direction and location of the crack obtained by the boundary element analysis was consistent with those observed in the frit salvage procedure. The panel corner pullout protector has been attached on the outer heel radius's diagonal corner of the panel to prevent thermal failure along the diagonal corner. As a result, the maximum stresses on the diagonal corner have been decreased significantly while the stress of point A at the seal edge has increased tremendously. Consequently, the crack continues to transverse along the high stress lines at the seal edge as required for a successful frit salvage process.

5 Boundary Element Technology XII 465 REFERENCES 1. Reddy, J. N., An Introduction to the Finite Element Method. 1st Edition, McGraw Hill, New York, Brebbia, C. A., Telles, J. C. F. and Wrobel, L C., Boundary Element Techniques. 1st Edition, Springer-Verlag, Berlin, Trevelyan J., Boundary Elements for Engineers: Theory and Application. Computational Mechanics Publications, Southampton, UK, Cha, C. Y., Iyer, S. and Trevelyan J., "Stress Failure Analysis of Neck Breakage of Cathode Ray Tube," Proceedings of 17th International Conference on Boundary Element Methods in Engineering, Vol. 1, Computational Mechanics Publications, Southampton, UK, pp , Cha, C. Y., Iyer, S. and Trevelyan J., "Comparison of Convergence and Modeling Times for Cathode Ray Tube Stress Analysis with the Finite Element and Boundary Element Methods," Boundary Element Communications, Computational Mechanics Publications, Southampton, UK, Vol. 6, No. 1, pp. 1-5, January, Aliabadi, M. H. and Rooke, D. P. Numerical Fracture Mechanics, Computational Mechanics Publications, Southampton, UK, and Kluwer Academics Publisher, Dordrecht, BEASY, Computational Mechanics, Southampton, UK, and Billerica, MA.

6 466 Boundary Element Technology XII Table 1. Principal Stress Comparison of the Outer Panels Between Salvage Tube and Salvage Tube with the Panel Corner Pullout Protector. POINTS Mesh No. The stress of salvage tube (psi) The stress of salvage tube with the panel corner pullout protector (psi) The amount of stress changed (psi) A B C D E THE THERMAL STRESS ANALYSIS OF 27V SALVAGE TUBE Figure 1 Main Features of 27V Tube with 3D Boundary Element Model

7 Boundary Element Technology XII 467 Actual Crack Line Figure 2a The Detailed Descriptions of Panel The Panel Corner Pullout's Connector The Panel Pullout Protector Figure 2b: The Panel Corner Pullout Protector with Panel

8 468 Boundary Element Technology XII TEMPERATURE SOLUT.IQNS Of 27V SALVAGE TUBE IH E+03 HI E E+03 ' 1J E E E E E E E+02 Temperature MAX = E+03 MIN = E+02 Figure 3: Temperature Solutions of Outer Glass Surface TEMPERATURE SOLUTIONS OF. 27V SALVAGE TUBE E E+03 Temperature MAX E+03 MIN = E+02 Figure 4 Temperature Solutions of Inner Glass Surface

9 THE THERMAL STRESS ANALYSIS OF 27V SALVAGE TUBE Boundary Element Technology XII e eH) E E E E E E E E E+02 Principal stress (max-alg) MAX= E+04 MIN= E+02 Figure 5 Principal Stress Solution of Outer Glass Surface E E-KW 0.209G7E E EKW EKW E E E-KB Rmcipal stress (max-alg) MAX E-KW MG^ E+02 Figure 6 Principal Stress Solution of Inner Glass Surface

10 470 Boundary Element Technology XII THE THERMAL STRESS ANALYSIS OF 27V SALVAGE TUBE Figure 7 Principal Stresses of Outer Panel Surface E E+04 f E+04 I" E+04 r E * E+04 I E E+04 _ E+03 i E+03 [ E+02 Principal stress (max - alg) MAX = E+04 MIN = E+02 THE THERMAL STRESS ANALYSIS OF 27V SALVAGE TUBE Figure 8 The Principal Stress Directions of Outer Glass Surface

11 Boundary Element Technology XII 471 Load set 3 TEMPERATURE SOLUTIONS OF FILE NAME : PAT " THE CORNER PULLOUr- if PROTECTOR I3392E E E+03 II378E E E E E E E E E+02 MAX= E+03 MIN= E-H02 Figure 9 Temperature Solutions of Outer Glass Surface with the Panel Corner Pullout Protector Load act TEMPERATURE FILE NAME : CORNER PULLCU1 'V KflOTECTO? E E E E E E E E E E E E+02 Temperature JL, MAX = E+03 MIN = E+02 Figure 10 Temperature Solutions of Inner Glass Surface with the Panel Corner Pullout Protector

12 472 Boundary Element Technology XII JL E E E E E4O E E E E E E E-K)2 Principal slrtn (max - alg) MAX= E+04 MIN= E+02 Figure 11 Principal Stress Solutions of Outer Glass Surface with the Panel Corner Pullout Protector -JL, Principal slros (max - ak) MAX = 0.49W3E-HM MN E E E E-MM E-M) E E E E E E E E-K)2 Figure 12 Principal Stress Solutions of Inner Glass Surface with the Panel Corner Pullout Protector

13 Boundary Element Technology XII E E E E-KM E E E4O E-MM E E E E+02 Principal ttn** (max - ilg) MAX = EH-04 VON = E+02 Figure 13 Principal Stress of Outer Panel Surface with the Panel Corner Pullout Protector uwo?*' > uyrnoi ic. rcou^t tmcavn** THE. HESMAL 5T%3S WAL,' OF 27V S&.VAGE HJPF U"TH THl COM^H FILE WOE ', PATSI5,TUI \ ^MZtzm^^^'f:;.^- --4._t ^f-^fj«^??-*"? *2-^-i.. i * ^.I^^mfflfrrTiJtl-^f^" ijl'^'hfsfejl}!:!:^ Figure 14 Principal Stress Directions of Outer Glass Surface with the Panel Corner Pullout Protector