bsdhr
Aerospace
- Aug 4, 2006
- 12
Folks;
We have a situation in our hands, whereby a batch of parts have been processed that have incorrect metallurgical properties. And it is my attempt to use FEA to model this discrepancy.
The part is a hollow tube, machined from a solid bar of stainless steel. As part of manufacturing process, this part is cored eveloped to approx. 30 HRC prior to be induction hardened to approx. 51 HRC. This process is well controlled and repeatable. However, their is an inspection step prior to core developing or rather machining the bar. And it is to check the impurities in the material. What our metalurgical inspector does is cuts a cross section of incoming bar and checks hardness as well as impurities such are voids, inclusions etc. If he had done his job, this batch of wrong parts would have been avoided. Because what he overlooked and now a problem, is the fact that their are strands of hard material buried uniformly within rest of the material ( see attached link, light color represents strands, dark core-developed material). The cut sections at cross-section of the bar would look like dispersion of spots all over the cross-sectional area that represent these local hard material strands. Another cut section along the length of the tube (shown in picture) would show these same spots running along the length of tube. If you put two cross-sections together it gives an appearance of chopped fibres in composite material.
Now the problem is when the part was designed, uniform and homogeneous material was assumed. Now with this non-homogeneous material configuration, the question is "what is the effect on yield and/ or fatigure life?".
Their is no literature that I am aware of answering this issue, so were thinking of what if we simulate this hardness variation within a material as composite material. I am thinking about chopped fibres, aligned lengthwise but disperesed in an epoxy. In this case, spoxy would be steel material with uniform-homogeneous material property, while the fibres would represent hardened strands. Next step would be to measure quantatively to come up with the volume of strand in given volum of total material. This can be done by taking some representative area and measuring the lengths of strands and their approx. cross-sectional areas.
Once this representative element is developed, we conduct simple pull-test to measure tensile strength and use that as ultimate strength of the material for the part.
It sounds like pretty involved solution process, but not sure if this right approach. Please share you thoughts on this. Thanks!
We have a situation in our hands, whereby a batch of parts have been processed that have incorrect metallurgical properties. And it is my attempt to use FEA to model this discrepancy.
The part is a hollow tube, machined from a solid bar of stainless steel. As part of manufacturing process, this part is cored eveloped to approx. 30 HRC prior to be induction hardened to approx. 51 HRC. This process is well controlled and repeatable. However, their is an inspection step prior to core developing or rather machining the bar. And it is to check the impurities in the material. What our metalurgical inspector does is cuts a cross section of incoming bar and checks hardness as well as impurities such are voids, inclusions etc. If he had done his job, this batch of wrong parts would have been avoided. Because what he overlooked and now a problem, is the fact that their are strands of hard material buried uniformly within rest of the material ( see attached link, light color represents strands, dark core-developed material). The cut sections at cross-section of the bar would look like dispersion of spots all over the cross-sectional area that represent these local hard material strands. Another cut section along the length of the tube (shown in picture) would show these same spots running along the length of tube. If you put two cross-sections together it gives an appearance of chopped fibres in composite material.
Now the problem is when the part was designed, uniform and homogeneous material was assumed. Now with this non-homogeneous material configuration, the question is "what is the effect on yield and/ or fatigure life?".
Their is no literature that I am aware of answering this issue, so were thinking of what if we simulate this hardness variation within a material as composite material. I am thinking about chopped fibres, aligned lengthwise but disperesed in an epoxy. In this case, spoxy would be steel material with uniform-homogeneous material property, while the fibres would represent hardened strands. Next step would be to measure quantatively to come up with the volume of strand in given volum of total material. This can be done by taking some representative area and measuring the lengths of strands and their approx. cross-sectional areas.
Once this representative element is developed, we conduct simple pull-test to measure tensile strength and use that as ultimate strength of the material for the part.
It sounds like pretty involved solution process, but not sure if this right approach. Please share you thoughts on this. Thanks!