Talk:Linear elasticity

Physicists/engineers notation
How's about having the standard 'physicists/engineers' notation for these equations included also. I doubt that many of these over 40 know the mathematicians 'differential forms' notation. Linuxlad 9 July 2005 12:12 (UTC)


 * I added a page called 3-D Elasticity that does this, I believe. I don't know what everyone wants to do with it, but I would have no problem combining the two pages.  I actually didn't see this one until after I'd made the 3-D elasticity one.  I think my version is at least as useful, since I don't see how a person could get any meaning from the Einsteinian notation forms, but then I'm not a mathmatician. - EndingPop 21:00, 23 December 2005 (UTC)
 * The linear elasticity article should be a long article incorporating both the engineers point of view, and then at the end, possibly the mathematical condensed presentation. I think the contents of the 3-D Elasticity article should be merged into this one. Berland 07:56, 20 February 2007 (UTC)

Fixed equation
The equation for the acoustic algebraic operator was obviously wrong, since the unit tensor and the k-k dyad don't have the same dimensions. I think it is fixed, but I am not sure about the eigenvalue statements.

With regard to the Einstein notation, yes the meaning is a little more obscure, but it is a compact notation, and it carries an automatic error check as to whether an equation has physical meaning: Both sides of an equation must have matching indices, and any quantity which has all indices in pairs is a quantity which is physically real, and does not depend on the coordinate system. Many theorems which are laborious to prove without the index notation can be quickly proved using the index notation and its rules. I wrote the indexed version of the acoustical operator, and you can see how neat it is. PAR 18:41, 6 May 2006 (UTC)
 * That's wonderful when you're proving something, but that doesn't discount the fact that it hides the meaning. For an encyclopedia, being able to understand it is far more important than the mathematical elegance that may be inherent in the notation. - EndingPop 21:36, 26 June 2007 (UTC)

Merge
I put up merging templates. The elastic wave article can be merged in, but might require some work on this article. The 3-D elasticity requires quite a bit of effort to be successfully merged in here, but it should happen some time anyway. --Berland 19:07, 26 June 2007 (UTC)

I took the liberty of starting the merging of the 3d-Elasticity article with the Linear Elasticity article. The former uses elements of the latter, which is a more general topic that covers equilibrium equations, constitutive relations, compatibility equations and boundary conditions. A lot of work (reorganizing the Table of Contents, history of theory of elasticity, mathematical formulations, etc) needs to be done. Comments...Sanpaz (talk) 23:27, 8 January 2008 (UTC)


 * A few articles are related to this Linear Elasticity article. These are Elasticity (physics), stress (physics), Hooke's Law, Strain (materials science), Strain tensor, Piola-Kirchhoff stress tensor (any others ????). Somehow, all these articles need to be merged into, or referenced (integrated) from, the Linear elasticity article. There is duplication and a lack of context with the current state of all articles. Sanpaz (talk) 23:56, 8 January 2008 (UTC)


 * I agree completely. I also agree with the move you have just made. I think if you go ahead, make changes slowly, so people can ponder them, that this would be a great improvement. Also, I don't know what to do about the simple notation vs. vector notation vs. Einstein notation problem. I prefer the Einstein notation usually ($$\partial_i V_i$$), the vector notation ($$\nabla \cdot \mathbf{V}$$) sometimes. The simple notation


 * $$\frac{\partial V_x}{\partial x}+\frac{\partial V_y}{\partial y}+\frac{\partial V_z}{\partial z}$$


 * seems to be just a needless blizzard of symbols. On the other hand, before I learned Einstein notation, it would have been the most obvious and preferable statement. It would be nice to speak to both groups but writing every equation three different ways does not seem like a good idea. PAR (talk) 00:19, 9 January 2008 (UTC)


 * you took the words out of my mouth ;). I was thinking the same thing about a the consistency of notations. Index notation is the most elegant a suppose. However, most engineers learn with the engineering notation (simple notation). So, somehow it would be nice to have both for more clarity and usefulness for more people. Sanpaz (talk) 00:31, 9 January 2008 (UTC)


 * Would it make sense to have two articles? Two articles that sort of "cooperated" with each other. That was the idea behind having "3-D elasticity" and "Linear elasticity" as separate articles, but they never really cooperated. PAR (talk) 01:15, 9 January 2008 (UTC)


 * My two cents worth: I was looking through these articles for the first time a few weeks ago to see what to reference from an article I was writing on Lamb waves, and I really liked 3-D Elasticity both for its style and for the level at which the math was presented. The most compact math formulations may be the most elegant, but they are also the most inscrutable to the common reader who we presumably hope to be educating with these articles. I'm just a regular Ph.D. physicist who uses these elasticity concepts for a living and uses Wikipedia for information. When I'm teaching this stuff to professional engineers who forgot their calculus, I show the compact form of the 3-D wave equation to the class to impress them how elegant and simple it really is, and then I give a layman's explanation of what the terms represent because they don't know a del from a hole in the ground. Of course, one man's meat is another man's poison. Good work and good luck guys! Adrian Pollock (talk) 00:48, 18 January 2008 (UTC)
 * I agree with you, the article needs also the engineering notation. But, first is better to organize the whole structure of the article, then people can start later inserting the different notations. Sanpaz (talk) 05:58, 18 January 2008 (UTC)

I don't agree with all of the proposed mergers. Definitely Elasticity (physics) should be a separate article from Linear elasticity; the former is a more general article that includes nonlinear elasticity. Nor do I see a problem with having separate articles on Strain tensor and Stress (physics), though Strain (materials science) does look like an unnecessary duplication. Also it makes sense to have a separate article on Hooke's Law, though it should be devoted to the one dimensional linear case (springs)--there should be some places in Wikipedia for elementary physics. Perturbationist (talk) 04:59, 14 January 2008 (UTC)


 * I guess I did not explained correctly the changes I was suggesting. My apologies for the misunderstanding. I actually agree with most of what you just said Perturbationist. I did not mean to suggest not having a strain and stress articles. My suggestion was to keep the concepts related to each of those topics in their respective article, and just reference the equations in the Linear Elasticity article if needed. And the same with other related articles. However, the Hooke's Law article should not be devoted only to springs. The more in depth formulation should be in it. But, for usefulness to 'elementary physics' it should have the spring theory, like it has already. Also a strongly agree about the Elasticity article being the one with higher hierarchy, with Linear Elasticty, Non Linear Elasticity, Viscoelasticity, etc, starting from there. Please let me know what you think Sanpaz (talk) 06:05, 14 January 2008 (UTC)


 * Oh, now that I see what you're proposing, I really like your idea. Perturbationist (talk) 03:25, 16 January 2008 (UTC)

Introductory Paragraph
I made some changes (pretty major ones) to the introductory paragraph. Let me know if I went overboard on my edit. Gpayette 06:49, 10 October 2007 (UTC)

Major changes to this article
I made a lot of changes to this article, and I would like to explain the reason for these, and how I see this article shaping out. The article should first state what Linear Elasticity is, which the first paragraph has done already (some changes in the way is written may be necessary). Later the article has a section that explains the boundary value problems that the theory of elasticity deals with and the way they are solved, showing the equations that have to be considered (no derivations). Once the theory of elasticity has been addressed as a whole, more detail can be added in other sections that should be developed, e.g. derivation of the Navier-Cauchy and Beltrami-Michell equations, derivation of the 2D formulation for Plane strain and Plane stress cases, the Airy Stress Function. Because I deleted some sections related to strain, compatibility equations, and I will delete the Equilibrium equation sections, these ought to be addressed fully in other articles: Strain and the compatibility equations will be addressed in the articleStrain (materials science); and the Equilibrium equation (equation of motion) should be addressed in a new article or the article called Equation of motion (I'm not sure what is the best place). I will re-use the content deleted. I would like to know your comments Sanpaz (talk) 07:12, 13 January 2008 (UTC)


 * Sounds good - just as long as no material is deleted without making it very clear that it was deleted, and why. I think the edits made so far are definitely an improvement and the proposed redistribution of material makes sense. Thanks for doing it piecemeal so that each change can be looked at in isolation, rather than having to sort out a major overhaul. Also, regarding the constitutive equations at the top of the page, mention should be made that the equations after the first one are for the special case of an isotropic media (or that the elastic moduli are tensors ???). PAR (talk) 15:27, 13 January 2008 (UTC)


 * I included the Equation of Motion and the Equilibrium Equations in the article Equation of motion. Sanpaz (talk) 06:11, 14 January 2008 (UTC)

Sanpaz, as I said above I generally agree with what you're doing, and I would go so far as to say most of the material you removed does not need to be restored, anywhere. The section on equilibrium equations was overkill, as the article already points out that the equilibrium equations can be obtained from the equations of motion by setting the acceleration term to zero. There's really not much more to say (unless one wants to talk about solving the equilibrium equations, a nontrivial boundary value problem). The section that was titled "Strain-displacement equations" looked like nonsense. The one edit you made that I disagree with is removing the Constitutive equations section. It is true that these are called generalized Hooke's laws but that doesn't mean we should relegate them to the Hooke's law article. The constitutive equations, in their tensor form, are at the heart of what this subject is about. Even if they stay in the Hooke's law article we can still have a section here. A little duplication is not so bad. (Which is not to say the former Constitutive equations section could not stand improvement.) Perturbationist (talk) 04:02, 16 January 2008 (UTC)


 * I agree with that. The three equations are all fundamental to the concept of Linear elasticity and should not be broken up. PAR (talk) 04:53, 16 January 2008 (UTC)


 * Perturbationist and PAR, first of all, thanks for your comments. Regarding the Constitutive equations: the current article, as part of the "Elasticity of a homogeneous-isotropic body" section, has the constitutive equations in 3 different forms in index notation: one general, and two for an isotropic material as a function of stresses and strains. Perhaps what you are referring to is the absence of these equations in engineering notation, as it was in the original 3-D elasticity article of December 27, 2007. I think the important thing is to have in this article the equations (e.g. constitutive equations) that are necessary for the development of the elasticity theory. But, the details on how these equations came about should be in their respective article: in this case the Hooke's Law article should have the complete derivation in index notation, engineering notation, and perhaps in dyadic notation. Please let me know if I understood your question and if what I answered is logical.Sanpaz (talk) 06:34, 16 January 2008 (UTC)


 * Actually I overlooked the fact that Elasticity of a homogeneous-isotropic body already has the nonisotropic constitutive equations in it. As long as that stays in the article, I don't have a problem with your deletion of the old Constitutive equations section after all.  In fact the deleted section had a number of errors. Perturbationist (talk) 02:54, 17 January 2008 (UTC)


 * The last remark was too harsh. The deleted section is not that bad, and may be usable.  For the record, the original version is here.  Perturbationist (talk) 01:51, 18 January 2008 (UTC)


 * This one got me thinking. I suggest this option better: The intention of this section is for the development of the formulation of the elasticity problem for Isotropic materials (the beltrami-michell and Navier equations are for isotropic materials). Therefore, I will delete the "anisotropic" word, and put back "isotropic" in the title. Sanpaz (talk) 05:52, 17 January 2008 (UTC)


 * I'm not used to dealing with compatibility equations, but the statement that The kinematic equations ... include the strain-displacement equations ... and the compatibility equations does not seem right. The compatibility equations are not independent of the strain-displacement equations: they can be derived from the strain-displacement equations alone using only the commutativity of the derivative. Furthermore, the statement There are 81 compatibility equations, but only 6 are independent non-trivial equations, and from those 3 are actually independent needs to be fixed. What is the difference between "independent" and "actually independent"?


 * Also, please note that the "Navier-Cauchy" equations are identical to the "elastostatic equation" and we shouldn't give the impression (as the article does now) that they are somehow distinct.


 * And finally, the "displacement formulation" vs "stress formulation" discussion is in the wrong place. This discussion is for static situations only, and belongs in the static section.


 * Look, the article is about "Linear Elasticity". That means that the elasticity tensor is not assumed homogeneous and isotropic, at least to begin with. Maybe we should rename the article to "Isotropic Homogeneous Linear Elasticity" but we should at least introduce the elasticity tensor, and then note the simplifications resulting from the isotropic-homogeneous assumption. The main sections were (and should be, I think):


 * Introduction (the three main equations - no simplifications, include time dependence)
 * Isotropic-homogeneous simplifications
 * Isotropic-homogeneous elastostatic - no time dependence (Navier-Cauchy=elastostatic equation, and Beltrami-Michell. Equations derived from the main 3 - compatibility, biharmonic)
 * Isotropic-homogeneous elastodynamic - time dependence, wave equation
 * —Preceding unsigned comment added by PAR (talk • contribs) January 17, 2008


 * Sanpaz, the reason I removed the word isotropic from the first section is that it contains the general form of the constitutive equation $$ \sigma_{ij} = C_{ijkl} \, \varepsilon_{kl}$$. Hence it covers more than just the isotropic case. Along the lines of PAR's suggestion, I think we should say much more about the general case before diving into the isotropic case.   We should explain why there are only 21 independent components of $$C_{ijkl}$$, for example.
 * I somewhat answer your first statement here in the response to PAR below. However, the explanation about the 21 independent components should be developed in the Hooke's law article, and referenced in here. Sanpaz (talk) 03:04, 18 January 2008 (UTC)


 * PAR, you're right, there are a lot of things in this article that don't make sense. What's called the "strain-displacement equation" is really just the definition of the infinitesimal strain tensor.  It is not, by itself, a problem with "unknowns" to be "solved".    The Displacement Formulation and Stress Formulation paragraph seem to be talking about different sets of boundary conditions for the elastostatic problem, yet it's not clear why a change in boundary conditions would lead to a different PDE in the interior.  This article needs a lot of work. Perturbationist (talk) 02:29, 18 January 2008 (UTC)
 * Yes, this article needs quite a lot of work, but, that is the fun part. In elasticity one wants to find the stresses and strains inside a body that is being loaded or stressed. Therefore, it is a problem with unknowns to be solved. To answer you concer about the two formulations, the derivations should be included. They will.Sanpaz (talk) 03:04, 18 January 2008 (UTC)


 * I will answer PAR's comments one by one:
 * 1. The strain-displacement equations are Kinematic equations as they describe how the body moves or deforms. I included the compatibility equations as part of the kinematic equations as they relate the strains. (perhaps I am wrong on this last statement)
 * 2. Where does it say that the compatibility equations are independent of strain-displacements equations? You are right in your statement about the former being derived from the latter.
 * 3. There is an explanation about being only 3 independent equations. I will take the statement out until I back it up with the derivation.
 * 4. You are right about the article being for Linear Elasticity in general. My knowledge of elasticity is limited so far to the elastostatic case, so my view was a little bias. You just reminded me that the article also includes elastodynamics. So I agree with your sections suggestion. However, I do not understand the reason for the second section you mention. Wouldn't this be included in the next sections 3 and 4?
 * So, we should move the elastostatic formulations into its corresponding section as you suggested Sanpaz (talk) 02:23, 18 January 2008 (UTC)

I agree with Perturbationist, the strain displacement (SD) equations are simply a definition of the strain tensor. They are not kinematic, they do not describe how the body moves, only how it deforms. The compatibility equations are not part of the fundamental three equations, they are derived from the SD equation and should not be "tacked on" to the SD equation. They are not an alternative form of the SD equations. The compatibility equations follow from the SD equation but not vice versa - the SD equations do not follow from the compatibility equations. The SD equations are more fundamental, they carry more information.

I agree with Sanpaz that the sections are not the best. Maybe better would be

PAR (talk) 03:12, 18 January 2008 (UTC)
 * Introduction (3 fundamental equations)
 * Elasticity tensor - definition of homogeneous isotropic.
 * Homogeneous Isotropic elasticity tensor
 * Elastostatic (displacement and stress formulations, Navier-Cauchy/elastostatic and Beltrami-Michell, solutions, compatibility, biharmonic)
 * Elastodynamic
 * General elasticity tensor - a few words at least.
 * I like this new table of contents. It will be a matter of rearranging the article, and then continue from there. Sanpaz (talk) 05:18, 18 January 2008 (UTC)

I still think the strain tensor equation can be called a Kinematic set. Perhaps not kinematic "equations". The version of September of 2007 of this article had this approach. The main idea is that the elasticity problem has three categories (groups or sets) of equations: Statics (equilibrium equation - stresses and forces), kinematics (strain-displacement), and constitutive equations (hooke's law). Sanpaz (talk) 04:12, 18 January 2008 (UTC)


 * I looked up kinematics, it says Kinematics studies how the position of an object changes with time. So I have problems with the definition of the strain tensor being "kinematic". Also, the elasticity problem has three sets of equations: DYNAMICS (not statics), stress-strain, and constitutive equations. Statics is a special case of dynamics, in which time is not a factor. PAR (talk) 17:24, 18 January 2008 (UTC)
 * Point taken. I think the term Kinematics was used loosely in this case. Sanpaz (talk) 18:42, 18 January 2008 (UTC)


 * Its looking good, this article is starting to come together. Still have work to do, though. PAR (talk) 19:32, 18 January 2008 (UTC)


 * Talk to different people, and you get different definitions of kinematics. A dynamics expert will use "Kinematics is the study of the motion of bodies without respect to the forces that cause that motion". A solid mechanics person will say "Kinematics is a study of the deformation of a body without respect to the forces causing that deformation". Essentially, it's looking at a body's displacements (including rotations), whether those be rigid body motions or deformations, without attention to the causes of the displacements (forces). The strain-displacement equations are kinematic relations. - EndingPop (talk) 15:38, 21 January 2008 (UTC)


 * Well, I looked it up in Love (Treatise on the mathematical theory of elasticity) and in Marsden and Hughes (Mathematical foundations of Elasticity). Marsden and Hughes explicitly state that kinematics is the study of deformation, and Love implicitly states the same thing, so I stand corrected. Now I don't have a problem with the strain-displacement equations being called "kinematic equations" PAR (talk) 16:16, 21 January 2008 (UTC)


 * I've been away from the wiki for a while, and I'm surprised to see that 3-D elasticity went from being merged with this page to almost disappearing entirely. While I'm not opposed to all of the changes that have been made, I'm rather disappointed that the full equations were reduced to Einsteinian notation in all cases. This is a great way to save space when writing, but it obscures much of the physics in the equations. While some space has been saved on the page (a minor benefit), you've successfully excluded people from the audience for this page through obfuscation of the formulae in the name of brevity. - EndingPop (talk) 15:52, 21 January 2008 (UTC)


 * I agree that the full equations need to be written somewhere. I disagree that Einstein notation obscures the physics, it is most definitely not "obfuscation". It gives the same information as the full equations, and so it yields the same physical information. Einstein notation is a "language" that is actually richer than the full equations, but, like any language, it is gibberish to someone who has not learned it. If I don't understand French, that does not mean that someone speaking French is "obfuscating". I completely agree that the full equations need to be written somewhere, so that the article is immediately accessible to those who have not learned Einstein notation. The full equations were here a few days ago, and we need to restore them. These changes are mostly due to user Sanpaz. I would ask Sanpaz to stop editing for a while so that we can form a consensus. I hope you are willing to do a little fixing? PAR (talk) 16:16, 21 January 2008 (UTC)
 * EndingPop, I totally agree with you. The intent is not to get rid off the engineering notation. The idea is to have a single article which covers Linear elasticity in detail, and at the same time shows the equations in different notation. However, things do not get done overnight. The article needs more people to improve its quality and quantity.If you have time please help adding the equations in engineering notation. It is never a good idea to have 2 article talking about the same thing. The 3D-Elasticity is just a duplication of the topic of Linear elasticity. Let me know what you think. Sanpaz (talk) 16:19, 21 January 2008 (UTC)
 * I'll concede that "obfuscation" may not be the best word. I still believe that it cuts off some of the article's potential audience. My point is that an undergraduate can read a differential equation, and have at least a hope of understanding what it means. I wasn't fully exposed to index notation until Continuum Mechanics at the grad level, even though I knew the concepts behind linear elasticity well before that. To use your analogy, PAR, this page is exclusively for french speakers, despite a more global appeal of the content. I'm also fine with combining the pages, which I noted in the Talk:3-D elasticity page and the top of this one. My issue is that when the pages were first merged the information from both was still intact for the most part. Since then, the 3-D elasticity equations have disappeared or been changed to index notation, and I feel the old page was more deleted than merged at this point. Sure I'll help fix, I guess I need to know more about what the goal of this page is. Back before the merge this page seemed to be written with mathematicians in mind, while the other was written with engineers in mine (I know the latter to be true, since I wrote the 3-D page). - EndingPop (talk) 19:50, 23 January 2008 (UTC)
 * EndingPop, I am someone who is more familiar with tensor notation and the mathematical point of view, and before commenting on this article I was previously unfamiliar with the engineering notation. When I first read the 3-D elasticity article it seemed very strange and wrong to me, and I made some dismissive comments on this talk page about some of the material which was moved from there into this article which I now regret.  But one of the most rewarding things about Wikipedia is working with people from other backgrounds, and now that I am becoming familiar with the engineering point of view it makes more sense to me.  I agree with you that the content of the old 3-D article should be made available somewhere.  We should reconsider how to merge it into this page, though it's tricky.  Some of the concepts clash, for example the engineering notation has an incompatible definition of strain tensor.  Also, the reasoning in the 3-D article, based on counting equations and unknowns, is mathematically naive, though I understand its value as a heuristic.  I'm not sure what the solution is. Perturbationist (talk) 03:35, 25 January 2008 (UTC)
 * If you were to call me mathematically naive, I would not be offended. I'm used to leaving mathematical rigor behind to solve a problem, if the error introduced is small. Take mass lumping for solving explicit FEA as a good example of this. The incompatible versions of the strain tensor you mentioned, are you talking about the factor of 1/2 that engineers and mathematicians fight over? If not, please elaborate. - EndingPop (talk) 13:16, 25 January 2008 (UTC)


 * I think what Perturbationist might be talking about is the use of different symbols for different components of the stress tensor, e.g. &sigma; for the diagonal components, &tau; for the off-diagonal components. These components are not the "normal" and "shear" components of the tensor except in a particular coordinate system. Its like writing a vector as [Vx, Vy, Q] with a different symbol for the third component because it is the "up" component, which is different than the horizontal components. Writing a vector this way destroys the whole idea of a vector as a physical quantity which doesn't need a coordinate system in order to have meaning. The same holds true of the stress and strain tensors, they should not have different symbols for different components. PAR (talk) 19:29, 25 January 2008 (UTC)


 * I see your point, but if you specify which coordinate system is in use, does it still matter if you use τ? The normal and shear stresses (and strains) have physical meaning that is less obvious (avoiding the word "obscured") when not shown in what I would call a convenient coordinate system. On the topic of the factor of 1/2 that may or may not appear in different sources, I don't particularly care which is shown. That is, as long it's clear which is in use. - EndingPop (talk) 03:24, 31 January 2008 (UTC)


 * EndingPop, I don't believe you are naive, and I certainly meant no offense. I was simply pointing out that a system of partial differential equations does not necessarily have a solution just because there are an equal number of unknowns and equations.  As for the incompatibility of notation, I had in mind both the 1/2 in the shear components and the issue PAR pointed out.  I am not trying to argue that one way is "right" and the other is "wrong", I am just suggesting that mixing them both in a single article will be tricky. Perturbationist (talk) 02:46, 26 January 2008 (UTC)


 * It was a tongue in cheek comment, though I suppose that doesn't always come across in text. My reasoning for counting equations and unknowns is really just to show what's needed to fully define the problem. If you only have the strain-displacement equations, for example, you not only can't solve the problem, you're missing fundamental pieces. You're right, you can't always solve the system even if correctly posed. That's where FEM comes in, after all. - EndingPop (talk) 03:24, 31 January 2008 (UTC)

Equations (notation)
Trying to address the issue of including both the index notation and the engineering notation, I added in the article a table with both notations for the first equation (equation of motion). It is just a suggestion. If you guys think it is useful to present the rest of the article in this way, then we should do so. If you have a better solution or an improvement for this solution please comment. Sanpaz (talk) 21:15, 25 January 2008 (UTC)

Another option could be this

$$\ \sigma_{ij,j}+ F_i = \rho \partial_{tt} u_i \begin{cases} \frac{\partial \sigma_x}{\partial x} + \frac{\partial \tau_{yx}}{\partial y} + \frac{\partial \tau_{zx}}{\partial z} + F_x = \rho \frac{\partial u_x}{\partial t^2} \\

\frac{\partial \tau_{xy}}{\partial x} + \frac{\partial \sigma_y}{\partial y} + \frac{\partial \tau_{zy}}{\partial z} + F_y = \rho \frac{\partial u_y}{\partial t^2} \\

\frac{\partial \tau_{xz}}{\partial x} + \frac{\partial \tau_{yz}}{\partial y} + \frac{\partial \sigma_z}{\partial z} + F_z = \rho \frac{\partial u_z}{\partial t^2} \end{cases}$$

Sanpaz (talk) 21:20, 25 January 2008 (UTC)


 * The table looks nice. But I'm not sure we can do it through the whole article, and keep it readable. Perturbationist (talk) 02:53, 26 January 2008 (UTC)


 * I like the version shown on the page better than the one above, since it's clearer what is being shown. I agree that it'd be difficult to do on the whole page, but what about just on the major equations? - EndingPop (talk) 03:27, 31 January 2008 (UTC)


 * I think the major equations, yes. The more I work with this stuff, the more I realize that the index notation is the "language" of elasticity. Its most often (but not always) the best way to talk about the physics and mathematics of what is happening independent of any coordinate system effects. Its only when a particular solution in a particular coordinate system is sought that you go to the "full" equations. I think part of our job here is to communicate as much as possible to someone who has not learned the index notation, but also to motivate them to learn it, in order to more fully understand and speak about the physics of the situation without laboriously writing out every equation. So, yes, the major equations should be in all forms (full, abstract, and index), and perhaps a few demonstrations of the usefulness of the index notation for condensing the fully written equations without losing any meaning whatsoever, and in fact adding some insight. (e.g. the unmatched indices of a sum of terms must be the same and the matched indices must only occur in pairs, otherwise the equation is unphysical.) PAR (talk) 04:31, 31 January 2008 (UTC)

I Like the Collapsible Navs. I Think that with them the other equations can also be written using engineering notation. Sanpaz (talk) 16:18, 7 February 2008 (UTC)


 * Ok, good. I'll convert a few more. PAR (talk) 16:46, 7 February 2008 (UTC)

About Stub article St. Venenat equations
The link has changed from strain to this Saint Venant compatibility equations. I did not know of the existence of the second article. So my question is: Do you guys think this article should be merged into the strain article? I don't know if the compatibility equations have enough content to be its own article. I am going to suggest a merge. Sanpaz (talk) 15:47, 12 February 2008 (UTC)


 * I don't understand the Saint-Venant's compatibility condition article, but it seems pretty technical and might be too much to transfer to the strain article. I do know that the compatibility conditions have a lot of content to them and probably will eventually deserve their own article, but right now I don't have a preference either way. PAR (talk) 17:40, 12 February 2008 (UTC)

Lowercase delta symbol in Isotropic homogeneous media section
Would it be possible to define what the lowercase delta symbol in the Isotropic homogeneous media section represents? I've been going through the article and linked pages, including the Einstein summation conventions but am not seeing it anywhere. —Preceding unsigned comment added by 99.13.56.253 (talk) 10:44, 23 April 2010 (UTC)
 * done sanpaz (talk) 15:19, 23 April 2010 (UTC)

Boussinesq-Cerruti solution Error?
I think there may be an error in the Boussinesq-Cerruti solution. Specifically, the G_11 (and G_22) terms. What's there now is:

b/r+ x^2/r^3+a*(z/r/(r+z)-x^2/r/(r+z)^2) $$ but I think it should be

1/r+ x^2/r^3+a*(1/(r+z)-x^2/r/(r+z)^2). $$ I've left off the 1/(4*pi*mu) factor. My source is: JUNSHAN LI and EDWARD J. BERGER "A Boussinesq-Cerruti Solution Set for Constant and Linear Distribution of Normal and Tangential Load over a Triangular Area" Journal of Elasticity 63: 137-151, 2001. Also, I checked using Matlab's symbolic math toolbox, and the first formula does not satisfy the elastostatic equation. The second formula does. Also, if the first formula is right, a 1/r factors out of the entire tensor. This doesn't seem right. Thanks for a well-written article. —Preceding unsigned comment added by 134.174.9.30 (talk) 19:38, 29 September 2008 (UTC)
 * The equations have been corrected. There was a sign problem in the solution.  The solution given in Landau and Lifschitz p. 25 (eq. 8.18) has been checked by Maple and found to be correct.  That solution matches the new version of the equations with sign change. Bbanerje (talk) 03:02, 25 March 2009 (UTC)

Elasticity tensor
Elasticity tensor redirects here, but this article does not define what the elasticity tensor is. In fact, the article doesn't even mention "elasticity tensor", although this talk page certainly does. Is it the same thing as the stiffness tensor? This must be fixed. Mgnbar (talk) 18:48, 15 August 2010 (UTC)
 * Yes, the stiffness tensor is called the elasticity tensor by some physicists. The article redirects to Hooke's law now. Bbanerje (talk) 05:44, 17 August 2010 (UTC)

New section needed
I believe there is a need to classify, differentiate and link the following pages; Elastic wave, acoustic waves, acoustic emission, Rayleigh waves, Lamb waves, Love waves I believe this section should be add in elastodynamics under waves.Dr eng x (talk) 05:44, 3 September 2010 (UTC)

My understanding is elastic wave or stress wave is made up of P and S waves according to the equations. P is for pressure wave and S for shear waves. The waves on the surface are called Rayleigh waves. It is safe to say the surface depth is within the wave length. Guided waves (in layers) are either lamb or love waves.Dr eng x (talk) 06:36, 3 September 2010 (UTC)

Linear elasticity and the continuum hyphothesis
In the first paragraph it is claimed that linear elasticity relies on the hypothesis that there is no set whose cardinality is strictly between that of the integers and that of the real numbers. This is surprising and should be elaborated. —Preceding unsigned comment added by 83.255.40.174 (talk) 00:31, 15 September 2010 (UTC)


 * It doesn't rely on the continuum hypothesis; the author must have meant that it models the material as a continuum. I've fixed it. Mgnbar (talk) 12:53, 16 September 2010 (UTC)

Section: Comparison of Navier-Stokes and Navier-Cauchy
The article has a section "Comparison of Navier-Stokes and Navier-Cauchy" that is not ready for the public. In particular, it doesn't have any sentences. So I'm removing it from the article and storing it here, until someone fixes it up. Mgnbar (talk) 15:41, 16 February 2011 (UTC)


 * Thanks. Navier-Cauchy and Navier-Stokes are both ways of describing a continuum.  Dropping the fancy math, we can see how they can be described using the analogy of an image.  It could eventually make for a simple interactive flash demo, but I don't know if Wiki supports flash. - bmunden


 * Hi, bmunden. I don't know about Flash either. Videos and images are possible, although they're not interactive. Heck, just a finished version of the text below would be good for the article. Regards, Mgnbar (talk) 14:47, 2 March 2011 (UTC)

Condensed Navier-Cauchy equation:
 * $${\nabla}\cdot{\mathsf{C}:{\tfrac{1}{2} \left[{\nabla}\mathbf{u}+({\nabla}\mathbf{u})^T\right]\,\!}} + \mathbf{F} = \rho~\ddot{\mathbf{u}}$$

Navier-Stokes equation:
 * $$ \rho \left(\frac{\partial \mathbf{v}}{\partial t} + \mathbf{v} \cdot \nabla \mathbf{v} \right) = -\nabla p + \nabla \cdot\mathbb{T} + \mathbf{F}$$

Symmetry of the elasticity tensor error?
I believe the reasoning for the symmetry of the elasticity tensor was incorrect (under Anisotropic homogeneous media):

"As shown, the matrix $$ C_{\alpha \beta}\,\!$$ is symmetric, because of the linear relation between stress and strain. Hence, there are at most 21 different elements of $$ C_{\alpha \beta}\,\!$$."

A linear relation between stress and strain gives that $$ \sigma = C : \varepsilon\!$$ but it doesn't mean that the tensor $$C$$ is symmetric in $$\sigma$$ and $$\varepsilon$$. The symmetry comes from the energy function which I've now changed. — Preceding unsigned comment added by AndrewDaleCramer (talk • contribs) 06:36, 4 April 2013 (UTC)

Common Measurement/Examples?
I ended up in this article while trying to find out if there is a standard, or at least most common, unit for measuring elasticity and what that might be. However, this article appears to only include equations for modelling elasticity, but doesn't go into what you might produce your answer in, or need to convert it into.

Sorry if I've completely missed a more relevant article, but I think this particular article could do with some examples (possibly in an examples section, to keep the rest clear for use as reference material). For my own specific case I really just want some means of quickly describing a material's elasticity with numbers (as opposed to just using terms like "very" =), but yeah, some examples would improve the article I think, if anyone is willing to add a couple of common ones? -- Haravikk (talk) 21:39, 10 June 2013 (UTC)

Elastic momentum density and the second constitutive relation
The mathematical formulation on the page is correct for the vast majority of problems encountered in linear elasticity, but implicitly conceals a second constitutive relation. At present, the derivation in the article for linear elasticity starts with the equation of motion
 * $$\boldsymbol{\nabla}\cdot\boldsymbol{\sigma} + \mathbf{F} = \rho\ddot{\mathbf{u}}, $$

however this equation arises from considering the conservation of momentum equation
 * $$\boldsymbol{\nabla}\cdot\boldsymbol{\sigma} + \mathbf{F} = \dot{\mathbf{p}}, $$

where $$\mathbf{p}$$ denotes the momentum density. A constitutive relation for the momentum density in the form
 * $$ \mathbf{p} = \rho \dot{\mathbf{u}},$$

expands the field in terms of the velocity, and we arrive at the equation of motion. The net effect is that there are two constitutive relations in linear elasticity, which comprise Hooke's law and the momentum-velocity law
 * $$ \begin{align}

\boldsymbol{\sigma} &= \mathsf{C}:\boldsymbol{\varepsilon}, \\ \mathbf{p} &= \rho \dot{\mathbf{u}}. \end{align} $$ I appreciate that this is perhaps too subtle a point for the majority of readers but would there be interest in highlighting this somewhere on the page? The derivation would then be analogous to Maxwell's equations where there are two constitutive relations (such as $$\mathbf{D} = \varepsilon \mathbf{E}$$ and $$\mathbf{B} = \mu \mathbf{H}$$), it would establish a connection to relativistic mechanics, and also admit generalisations to Willis coupling which is an analogue to bi-isotropic materials in Maxwell's equations. — Preceding unsigned comment added by MJASmith (talk • contribs) 17:33, 5 January 2021 (UTC)


 * I agree wholeheartedly that the equation of motion should be written in terms of p_dot, even though I dont it matters for the differential formulation under the Lagrangian viewpoint (since we follow the material particles, as opposed to the Eulerian viewpoint). I've seen questions on misc. forum pages where some contributors bash Newtonian mechanics because they have the misconseption that F=ma is the basic statement of the second law, so this would be a step in the right direction.
 * About terming the momentum--velocity relation as constitutive, Im less certain about. I think that should be debated in relation to the article on constitutive relations, as I could not find this relation on that page.
 * Also, I think EM and mechanics are so different in nature that it shouldnt be a goal to align the two in that manner. 2A01:799:952:4500:D5C3:6421:BBA8:44CF (talk) 17:42, 6 February 2024 (UTC)