Stress-Strain Curve Approximation: Ramberg-Osgood Relationship

In structural analysis, particularly when components endure high loads exceeding their yield strength, accurate stress-strain curves are indispensable for precise calculations. However, obtaining these curves can pose challenges, often requiring approximations in the absence of readily available data. This blog post delves into one method of stress-strain curve approximation that closely mirrors reality.

Table of Contents

    Exploring Curve Simplifications

    Before delving into the preferred method, let’s explore existing simplifications commonly used when precise data is unavailable. These include the Perfectly Plastic Rigid Curve, Perfectly Plastic Elastic Curve, and Elastic with Linear Strain Hardening Curve.

    Perfectly Plastic Rigid Curve
    Perfectly Plastic Elastic Curve
    Elastic with Linear Strain Hardening Curve
    1. Perfectly Plastic Rigid: This model assumes a rigid tensile specimen that experiences no elastic deformation until it reaches its yield strength. Once the yield strength is attained, the material undergoes plastic strain at a constant stress level, without any strain hardening effect.
    1. Perfectly Plastic Elastic: In this simplification, stress and strain maintain a linear relationship within the elastic region until the material reaches its yield point. Beyond this threshold, immediate plastic deformation occurs at a constant stress level, without any strain hardening.
    1. Elastic with Linear Strain Hardening: This model depicts a material exhibiting initial elastic behavior, characterized by a linear stress-strain relationship. Upon surpassing the yield point, plastic deformation commences, accompanied by strain hardening, which also follow a linear stress-strain relationship.

    Introducing the Ramberg–Osgood Relationship

    The Ramberg–Osgood equation was created to describe the nonlinear relationship between stress and strain — that is, the stress–strain curve — in materials after their yield points. It is especially applicable to ductile metals that harden with plastic deformation, showing a smooth elastic-plastic transition.

    The equation is expressed as:

    ramberg osgood relationship

    Here, ε represents total strain, σ signifies stress, E stands for Young’s Modulus, while K and n are material-dependent constants.

    In the above expression, the term σ/E is the elastic strain portion and the term K(σ/E)n is the plastic strain portion.

    It’s crucial to note that this equation relates engineering stress with engineering strain, necessitating a transformation to true stress and true strain before utilization in Finite Element Analysis (FEA) software.

    The H. N. Hill Equation

    While the Ramberg–Osgood relationship offers precision, determining the parameters K and n can be challenging. To address this, H. N. Hill proposed an adaptation of the equation, simplifying parameter estimation by incorporating yield strength directly.

    The H. N. Hill Equation is expressed as:

    The H. N. Hill Equation

    Where σYS denotes yield strength.

    Notice in the expression above that the elastic strain component, represented by σ/E, remains unchanged. However, the plastic strain component, given by 0.002(σ/σYS)n, has undergone alteration, resulting in the absence of the term “K.” Instead, the yield strength is now present, which is a readily accessible material property.

    Let’s revisit the definition of yield strength. Yield strength denotes the maximum stress a material can sustain without experiencing considerable permanent deformation, typically characterized by a plastic strain of 0.2%. This parameter is crucial in indicating the threshold at which a material shifts from elastic to plastic behavior under tensile or compressive loading. Once the yield strength is surpassed, the material undergoes plastic deformation (even though there’s already 0.2% plastic strain, we usually set it to zero), indicating that it will not revert to its original shape after the removal of the load. It’s important to emphasize that the proportional limit is actually the stress the ends the elastic region and that presents zero plastic strain.

    stress strain curve

    In the H. N. Hill Equation, this definition is considered: when stress equals the yield strength, the corresponding plastic strain is 0.2%. Consequently, the stress at which the plastic strain is zero will be lower than the yield strength. This stress is known as the proportional limit.

    Calculating the Parameter “n”

    Determining the parameter “n” is crucial for curve accuracy. Utilizing known mechanical properties like Ultimate Tensile Strength (σUTS) and Maximum Elongation (εmax), “n” can be calculated using the equation:

    calculation of n

    Isolating “n”:

    expression that calculates material parameter n

    Converting to True Stress-Strain

    As engineering stress-strain curves are commonly used, conversion to true stress-strain curves is essential for FEA software. The conversion is achieved through the equations:

    true strain and stress expressions

    Example with ASTM A36

    Let’s delve into a step-by-step process using ASTM A36 as an example, one of the most common materials. The following mechanical properties can be easily found on the internet:

    ASTM A36 Mechanical Properties

    Step 1: Calculate “n”

    To begin, we calculate “n” using known mechanical properties:

    calculation of parameter n for ASTM A36

    Step 2: Derive the Engineering Stress/Strain Curve

    Utilizing the derived “n,” we obtain the engineering stress-strain curve:

    engineering stress strain expression
    engineering stress strain table
    engineering stress strain curve

    Step 3: Conversion to True Stress/Strain

    Converting engineering values to true stress/strain involves:

    true strain and stress expressions

    For example, with engineering stress of 300 MPa and strain of 0.01338960 from the table:

    true strain and stress calculation example

    We can compute these values for all data points in the table above. By superimposing these two charts, we can visually discern the disparity between engineering and true stress/strain. The resulting comparison is illustrated below:

    comparison between engineering and true stress/strain curves

    Notice that at small strains, the two curves closely resemble each other. However, as strain increases, the disparity between them becomes significant. This occurs because true stress calculation takes into account the instantaneous cross-sectional area, whereas engineering strain only considers the initial cross-sectional area. When a body undergoes deformation from a tensile load, the cross-sectional area decreases due to the Poisson’s effect. As stress is defined as force divided by area, a decrease in area results in higher stress. Consequently, true stress exceeds engineering stress for large strains.

    The new table generated for true stress/strain is:

    true stress versus true strain table

    Interested in delving deeper into engineering and true stress/strain? Explore this insightful blog post on the topic by clicking here.

    Step 4: Extraction of True Stress vs. True Plastic Strain Curve

    We currently possess the True Stress vs Total Strain Curve. In a finite element method software, the appropriate curve to employ is the true stress versus true plastic strain curve. Consequently, we must exclude the true elastic strain component from the prior curve.

    true plastic strain calculation

    Let’s employ the expression provided, starting from the line on the table with a stress of 250.8 MPa (True Yield Strength) and a strain of 0.00324473, and continue to the final line, where the stress reaches 480 MPa and the strain is 0.18232156. The calculated result is presented below:

    true plastic strain versus true stress table

    Step 5: Calculate True Stress for Zero True Plastic Strain

    To ensure compatibility with FEA software, the first line of the table should feature a true stress value corresponding to zero true plastic strain. This value can be determined through interpolation using the first two points of the preceding table.

    stress calculation for zero plastic strain

    Thus, we obtain a revised table suitable for FEA software integration:

    true plastic strain versus true stress table for FEA

    Acknowledged, it’s important to note that the stress value for zero plastic strain does not align with the yield strength. Additionally, at a true plastic strain of 0.001995511 = 0.1995511% (second line of the table above), the stress equals 250.8 MPa, which matches the true yield strength (while the engineering yield strength remains at 250 MPa). The discrepancy in the plastic strain for yield strength arises because the strain was converted to true strain, resulting in a slightly different value from the nominal 0.2%.

    Important Note

    By employing this method to derive the true stress/strain curve, it’s feasible to incorporate up to 0.2% plastic strain into your model while still assuming that the part has not yielded. This assertion holds true because only at a true plastic strain of 0.001995511 = 0.1995511% (rounded to 0.2%) do we observe a stress value equal to the material’s yield strength.

    Explore Further with Python

    We’ve developed a Python script to handle these calculations for you, offering results with enhanced accuracy thanks to additional features. Here’s what sets it apart:

    • Optimal stress distribution is considered up to the ultimate tensile strength.
    • Beyond the ultimate tensile strength, the script establishes a linear function with a slope equal to 7% of the slope of the last two points. This conservative approach ensures safer calculations at higher stress levels. Without this adjustment, the software would simply rely on the slope of the last two points.

    Learn More

    Enhance your understanding of FEA with the following blog posts:

    References

    • Ramberg W and Osgood W. R. – Description of Stress-Strain Curves by Three Parameters – No. 9023, NACA, 1943
    • Hill H. N. – Determination of Stress-Strain Relations from “Offset” Yield Strength Values
    • Office of Aviation Research – DOT/FAA/AR-MMPDS-01 – Metallic Materials Properties Development and Standardization 
    Picture of Renato Carvalho

    Renato Carvalho

    I'm a CAE Durability Engineer specializing in structural analysis using the Finite Element Method (FEM). I perform simulations involving buckling, fatigue, misuse, dynamic explicit events, and strength assessments, using tools such as ANSA, ABAQUS, NASTRAN, PAM-CRASH, LS-DYNA, nCode, META Post, and Hyperview. I hold a specialization in Numerical Structural Analysis with FEM from ESSS and a Mechanical Engineering degree with a focus on the automotive industry from Centro Universitário FEI. I'm also experienced in the design of metallic structures and actively create technical content about FEM and structural engineering. Currently, I am pursuing a master’s degree at the Universidade Federal do ABC (UFABC), where I am developing a new user-defined finite element for bolted connections, intended for the design and analysis of complex steel structural connections.

    14 Comments.

    1. “Acknowledged, it’s important to note that the stress value for zero plastic strain does not align with the yield strength. Additionally, at a true plastic strain of 0.001995511 = 0.01995511%” <– Typo. 0.1995511%.

    2. Interesting article. I have a question though. Can you confirm that maximum elongation is comprised of elastic and plastic strain? MMDPS derives elongation as the permanent plastic strain measured after fracture of a tensile specimen. There is no elastic strain after fracture. Also the plastic strain is a combination of strain hardening up to ultimate tensile strength and necking strain. Using the assumption the strain at Ultimate is equal to elongation at failure would not be conservative as it contains the necking strain as well. I would be interested to hear your thoughts on this.

    3. in the formula of n you assume that the engineering UTS point is (A%/100=epsilon_max,UTS) but it isn’t in that way. Epsilon_max is the rupture elongation post necking, it isn’t the value corresponding at UTS point. You have to know the eps corresponding at UTS value of sigma to use the n formula. How can it do that? thanks

      1. Hi Albert. Great question. This is something I should have explained in more details. Using the Maximum Elongation at Break is not correct, but it is conservative, and that’s why we can use it. To understand why, let’s imagine a simpler scenario of a bilinear stress strain curve (first straight line for elastic regime, and second straight line connecting the yield point to the UTS point). When building this second line, if I use Maximum Elongation at Break strain for UTS (which is not correct), the inclination of this line is lesser, and therefore the stiffness is lesser because Maximum Elongation at Break is greater than the actual strain at UTS. If the inclination/stiffness is lesser, then you’re being conservative, so it’s Ok. If you have strain at UTS, then it’s better to use it instead of Maximum Elongation at Break, though.

        1. Isn t it the opposite? If i assume (A%/100,UTS) as that point, i assume that i arrive at UTS for a elongation bigger thant real true case. So in a simulation with a certain total deformation before necking i read a stress which is less than the real case, for the reason that inclination curve is lesser. Thanks for the replay

          1. Yes, the stress values before the UTS may be lower than the real stress for the same plastic strain level, but this does not mean the analysis is not conservative. In strength analyses where plastic deformation occurs, plastic strain is more important than stress. If I were writing a report for a component with plastic deformation, I would focus on plastic strain results, not on stress values.

            To explain this more clearly, imagine the following example. You perform a strength analysis of a component, and the acceptance limit is a plastic strain of 0.5% or less. First, you create the stress–strain curve using the correct strain at UTS. With the applied loads, the maximum plastic strain is 0.5%, so the component meets the acceptance limit.

            Now imagine you do not have the correct strain at UTS and instead use the maximum elongation at break to build the stress–strain curve. You run the same analysis with the same loads, and the maximum plastic strain increases to, for example, 0.7%. This result is conservative because it forces you to strengthen the component, rather than weakening it just to meet the acceptance limit.

    4. in the formula of n you assume that the engineering UTS point is (A%/100=epsilon_max,UTS) but it isn’t in that way. Epsilon_max is the rupture elongation post necking, it isn’t the value corresponding at UTS point. You have to know the eps corresponding at UTS value of sigma to use the n formula. How can it do that? thanks.

    5. Hi Renato, Could I do a question?
      where Do you find the “n” equation?
      There is a differen equation In Ramberg osgood document
      thanks

    Leave a Reply

    Your email address will not be published. Required fields are marked *

    Recommended For You