Tag Archives: residual stress

Star sequence minimises distortion

Leave a reply

It is some months since I have written about engineering so this post is focussed on some mechanical engineering.  The advent of pneumatic and electric torque wrenches has made it impossible for the ordinary motorist to change a wheel because it is very difficult to loosen wheel nuts by hand when they have been tightened by a powered wrench which most of us do not have available.  This has probably made motoring safer but also means we are more likely to need assistance when we have a flat tire.  It also means that the correct tightening pattern for nuts and bolts is less widely known.  A star-shaped sequence is optimum, i.e., if you have six bolts numbered sequentially around a circle then you start with #1, move across the diameter to #4, then to #2 followed by #5 across the diameter, then to #3 and across the diameter to #6.  This sequence is optimum for flanges, bolted joints in the frames of buildings and joining machine parts as well as wheel nuts.  We have recently discovered that it works in reverse, in the sense that it is the optimum sequence for releasing parts made by additive manufacturing (AM) from the baseplate of the AM machine (see ‘If you don’t succeed try and try again’ on September 29th, 2021).  Additive manufacturing induces large residual stresses as a consequence of the cycles of heat input to the part during manufacturing and some of these stresses are released when it is removed from the baseplate of the AM machine, which causes distortion of the part.  Together with a number of collaborators, I have been researching the most effective method of building thin flat plates using additive manufacturing (see ‘On flatness and roughness’ on January 19th, 2022).  We have found that building the plate vertically layer-by-layer works well when the plate is supported by buttresses on its edges.  We have used two in-plane buttresses and four out-of-plane buttresses, as shown in the photograph, to achieve parts that have comparable flatness to those made using traditional methods.  It turns out that optimum order for the removal of the buttresses is the same star sequence used for tightening bolts and it substantially reduces distortion of the plate compared to some other sequences.  Perhaps in retrospect, we should not be surprised by this result; however, hindsight is a wonderful thing.

The current research is funded jointly by the National Science Foundation (NSF) in the USA and the Engineering and Physical Sciences Research Council (EPSRC) in the UK and the project was described in ‘Slow start to an exciting new project on thermoacoustic response of AM metals’ on September 9th 2020.

Image: Photograph of a geometrically-reinforced thin plate (230 x 130 x 1.2 mm) built vertically layer-by-layer using the laser powder bed fusion process on a baseplate (shown removed from the AM machine) with the supporting buttresses in place.

Sources:

Patterson EA, Lambros J, Magana-Carranza R, Sutcliffe CJ. Residual stress effects during additive manufacturing of reinforced thin nickel–chromium plates. IJ Advanced Manufacturing Technology;123(5):1845-57, 2022.

Khanbolouki P, Magana-Carranza R, Sutcliffe C, Patterson E, Lambros J. In situ measurements and simulation of residual stresses and deformations in additively manufactured thin plates. IJ Advanced Manufacturing Technology; 132(7):4055-68, 2024.

If you don’t succeed, try and try again…

4 Replies

Photograph of S-shaped plateYou would not think it was difficult to build a thin flat metallic plate using a digital description of the plate and a Laser Powder Bed Fusion (L-PBF) machine which can build complex components, such as hip prostheses.  But it is.  As we have discovered since we started our research project on the thermoacoustic response of additively manufactured parts (see ‘Slow start to an exciting new project on thermoacoustic response of AM metals‘ on September 9th, 2020).  L-PBF involves using a laser beam to melt selected regions of a thin layer of metal powder spread over a flat bed.  The selected regions represent a cross-section of the desired three-dimensional component and repeating the process for each successive cross-section results in the additive building of the component as each layer solidifies.  And there in those last four words lies the problem because ‘as each layer solidifies’ the temperature distribution between the layers causes different levels of thermal expansion that results in strains being locked into our thin plates.  Our plates are too thin to build with their plane surfaces horizontal or perpendicular to the laser beam so instead we build them with their plane surface parallel to the laser beam, or vertical like a street sign.  In our early attempts, the residual stresses induced by the locked-in strains caused the plate to buckle into an S-shape before it was complete (see image).  We solved this problem by building buttresses at the edges of the plate.  However, when we remove the buttresses and detach the plate from the build platform, it buckles into a dome-shape.  Actually, you can press the centre of the plate and make it snap back and forth noisily.  While we are making progress in understanding the mechanisms at work, we have some way to go before we can confidently produce flat plates using additive manufacturing that we can use in comparisons with our earlier work on the performance of conventionally, or subtractively, manufactured plates subject to the thermoacoustic loading experienced by the skin of a hypersonic vehicle [see ‘Potential dynamic buckling in hypersonic vehicle skin‘ on July 1st 2020) or the containment walls in a fusion reactor.  Sometimes research is painfully slow but no one ever talks about it.  Maybe because we quickly forget the painful parts once we have a successful outcome to brag about. But it is often precisely the painful repetitions of “try and try again” that allow us to reach the bragging stage of a successful outcome.

The research is funded jointly by the National Science Foundation (NSF) in the USA and the Engineering and Physical Sciences Research Council (EPSRC) in the UK (see Grants on the Web).

References

Silva AS, Sebastian CM, Lambros J and Patterson EA, 2019. High temperature modal analysis of a non-uniformly heated rectangular plate: Experiments and simulations. J. Sound & Vibration, 443, pp.397-410.

Magana-Carranza R, Sutcliffe CJ, Patterson EA, 2021, The effect of processing parameters and material properties on residual forces induced in Laser Powder Bed Fusion (L-PBF). Additive Manufacturing. 46:102192

Poleidoscope (=polariscope + kaleidoscope)

1 Reply

A section from a photoelastic model of turbine disc with a single blade viewed in polarised light to reveal the stress distribution.Last month I wrote about the tedium of collecting data 35 years ago without digital instrumentation and how it led me to work on automation and digitalisation in experimental mechanics [see ‘35 years later and still working on a PhD thesis‘ on September 16th, 2020].  Thirty years ago, one of the leading methods for determining stresses in components was photoelasticity, which uses polarised light to generate fringe patterns in transparent components or models that correspond to the distribution of stress.  The photoelastic fringes can be analysed in a polariscope, of which the basic principles are explained in a note at the end of this post.  During my PhD, I took hundreds of black and white photographs in a polariscope using sheets of 4×5 film, which came in boxes of 25 sheets that you can still buy, and then scanned these negatives using a microdensitometer to digitise the position of the fringes.  About 15 years after my PhD, together with my collaborators, I patented the poleidoscope which is a combination of a polariscope and a kaleidoscope [US patents 6441972 & 5978087] that removes all of that tedium.  It uses the concept of the multi-faceted lens in a child’s kaleidoscope to create several polariscopes within a compound lens attached to a digital camera.  Each polariscope has different polarising elements such that photoelastic fringes are phase-shifted between the set of images generated by the multi-faceted lens.  The phase-shifted fringe patterns can be digitally processed to yield maps of stress much faster and more reliably than any other method.  Photoelastic stress analysis is no longer popular in mainstream engineering or experimental mechanics due to the simplicity and power of digital image correlation [see ‘256 shades of grey‘ on January 22nd, 2014]; however, the poleidoscope has found a market as an inspection device that provides real-time information on residual stresses in glass sheets and silicon wafers during their production.  In 2003, I took study leave for the summer to work with Jon Lesniak at Glass Photonics in Madison, Wisconsin on the commercialisation of the poleidoscope.  Subsequently, Glass Photonics have  sold more than 250 instruments worldwide.

For more information on the poleidoscope see: Lesniak JR, Zhang SJ & Patterson EA, The design and evaluation of the poleidoscope: a novel digital polariscope, Experimental Mechanics, 44(2):128-135, 2004

Note on the Basic principles of photoelasticity: At any point in a loaded component there is a stress acting in every direction. The directions in which the stresses have the maximum and minimum values for the point are known as principal directions. The corresponding stresses are known as maximum and minimum principal stresses. When polarised light enters a loaded transparent component, it is split into two beams. Both beams travel along the same path, but each vibrates along a principal direction and travels at a speed proportional to the associated principal stress. Consequently, the light emerges as two beams vibrating out of phase with one another which when combined produce an interference pattern.   The polarised light is produced by the polariser in the polariscope and the analyser performs the combination. The interference pattern is observed in the polariscope, and the fringes are contours of principal stress difference which are known as isochromatics. When plane polarised light is used black fringes known as isoclinics are superimposed on the isochromatic pattern. Isoclinics indicate points at which the principal directions are aligned to the polarising axes of the polariser and analyser.

Image: a section from a photoelastic model of turbine disc with a single blade viewed in polarised light to reveal the stress distribution.

Slow start to an exciting new project on thermoacoustic response of AM metals

3 Replies

We held the kick-off meeting for a new research project this week.  It’s a three-way collaboration involving three professors based in Portugal, the UK and USA [Chris Sutcliffe, John Lambros at UIUC and me]; so, our kick-off meeting should have involved at least two of us travelling to the laboratory of the third collaborator and spending some time brainstorming about the challenges that we have agreed to tackle over the next three years.  Instead we had a call via Skype and a rather procedural meeting in which we covered all of the issues without really engendering any excitement or sparking any new ideas.  It would appear that we need the stimulus of new environments to maximise our creativity and that we use body language as well as facial expressions to help us reach a friendly consensus on which  crazy ideas are worth pursuing and which should be quietly forgotten.

Our new research project has a long title: ‘Thermoacoustic response of Additively Manufactured metals: A multi-scale study from grain to component scales‘.  In simple terms, we are going to look at whether residual stresses could be designed to be beneficial to the performance of structural parts used in demanding environments such as those found in reusable spacecraft, hypersonic flight vehicles and breeder blankets in fusion reactors.  Residual stresses are often induced during the manufacture of parts and are usually detrimental to the performance of the part.  Our hypothesis is that in additive manufacturing, or 3D printing, we have sufficient control of the manufacture of the part that we can introduce ‘designer stresses’ which will improve the part’s performance in demanding environments.  The research is funded jointly by the National Science Foundation (NSF) in the USA and the Engineering and Physical Sciences Research Council (EPSRC) in the UK and is supported by The MTC and Renishaw plc; you can find out more at Grants on the Web. The research will be building on our recent research on ‘Potential dynamic buckling in hypersonic vehicle skin‘ [posted July 1st, 2020] and earlier work, see ‘Hot stuff‘ on September 13th, 2012.  While the demanding environment is not new to us, we will be using 3D printed parts for the first time instead of components made by conventional (subtractive) machining and taking them to higher temperatures.

The thumbnail shows measured modal shapes for a subtractively-manufactured plate subject to the three temperature regimes: room temperature (left), transverse heating of the centre of the plate (middle) and longitudinal heating on one edge (right) from Silva, A.S., Sebastian, C.M., Lambros, J. and Patterson, E.A., 2019. High temperature modal analysis of a non-uniformly heated rectangular plate: Experiments and simulations. J. Sound & Vibration, 443, pp.397-410.