Stress-Strength Analysis: Difference between revisions

From ReliaWiki
Jump to navigation Jump to search
No edit summary
 
(29 intermediate revisions by 5 users not shown)
Line 1: Line 1:
{{Stubs}}
{{Template:LDABOOK|Additional Reliability Analysis Tools|Stress-Strength Analysis}}
{{template:LDABOOK|25|Stress Strength Analysis}}
==Stress-Strength Analysis==


Using the same idea as illustrated in the previous section, one could determine the probability of failure based on the probability of stress exceeding strength. In this case, Eqn. (eqnStress) could be recast as:  
Stress-strength analysis has been used in mechanical component design. The probability of failure is based on the probability of stress exceeding strength. The following equation is used to calculate the expected probability of failure, ''F'':  


<br>
::<math>F=P[Stress\ge Strength]=\int_{0}^{\infty }{{{f}_{Strength}}(x)\cdot {{R}_{Stress}}(x)}dx\,\!</math>
::<math>P\left[ Stress\ge Strength \right]=\mathop{}_{0}^{\infty }{{f}_{Strength}}(x)\cdot {{R}_{Stress}}(x)\cdot dx</math>  


<br> In this case, the data for the strength set would be actual data that is indicative of the strength of the material (i.e. maximum applied stress to cause failure), and the stress data would be actual stress data of the material under use conditions. Weibull++'s Stress-Strength Wizard allows you to perform such calculations.
The expected probability of success or the expected reliability, ''R'' , is calculated as:


The Stress-Strength Calculator is accessible from the Tools folder of the Project Explorer. The tool first asks for the locations of the stress and strength data from among the folios available, as shown below.
::<math>R=P[Stress\le Strength]=\int_{0}^{\infty }{{{f}_{Stress}}(x)\cdot {{R}_{Strength}}(x)}dx\,\!</math>


[[Image:Firstchoice.jpg|thumb|center|400px| ]] 
The equations above assume that both stress and strength are in the positive domain. For general cases, the expected reliability can be calculated using the following equation:


Once the tool knows where to find data for stress and strength, it automatically calculates the reliability and graphs the distributions of the stress and strength data, as shown below.
::<math>R=P[{{X}_{1}}\le {{X}_{2}}]=\frac{1}{{{F}_{1}}(U)-{{F}_{1}}(L)}\int_{L}^{U}{{{f}_{1}}(x)\cdot {{R}_{2}}(x)}dx\,\!</math>


[[Image:Stress-strength.jpg|thumb|center|400px| ]] 
where:
:: <math>L\le {{X}_{1}}\le U\,\!</math>
::<math>\begin{align}
  & {{X}_{1}}:\text{ Stress } \\
& {{X}_{2}}:\text{ Strength } \\
\end{align}\,\!</math>
<br>
When U = infinity and L = 0, this equation becomes equal to previous equation (i.e., the equation for the expected reliability ''R'').  


Once the reliability has been calculated, the user has two different options for performing sensitivity analysis (seeing how the system performance changes with changes in parameter values). The user can manually alter the distribution parameters used to calculate the reliability by clicking the “Alter Parameters” link for the appropriate distribution, or the user can use the Parameter Estimator to find what parameters result in a specified target reliability. The “Alter Parameters” link allows the user to manually set what she would like the distribution parameter(s) to be. Alternatively, the Parameter Estimator utility allows the user to select a target reliability to reach and a distribution parameter to adjust. The utility then adjusts the parameter value until the specified target is reached and outputs the parameter value that does so, as shown below.  
==Confidence Intervals on the Probability==
Both the stress and strength distributions can be estimated from actual data or specified by engineers based on engineering knowledge or existing references. Based on the source of the distribution, there are two types of variation associated with the calculated probability: variation in the model parameters and variation in the probability values. Both are described next.


[[Image:Parameter estimator.jpg|thumb|center|300px| ]]
===Variation in Model Parameters===
If both the stress and strength distributions are estimated from data sets, then there are uncertainties associated with the estimated distribution parameters. These uncertainties will cause some degree of variation of the probability calculated from the stress-strength analysis. Therefore, we can use these uncertainties to estimate the confidence intervals on the calculated probability. To get the confidence intervals, we first calculate the variance of the reliability based on Taylor expansion by ignoring the 2nd order term. <math>R\,\!</math>, The approximation for the variance is:
 
::<math>Var\left[ R \right]=\int_{0}^{\infty }{Var\left[ {{f}_{1}}(x) \right]}{{\left[ {{R}_{2}}(x) \right]}^{2}}dx+{{\int_{0}^{\infty }{\left[ {{f}_{1}}(x) \right]}}^{2}}Var\left[ {{R}_{2}}(x) \right]dx\,\!</math>
 
Variance of <math>{{f}_{1}}(x)\,\!</math> and <math>{{R}_{2}}(x)\,\!</math> can be estimated from the Fisher Information Matrix. For details, please see [[Confidence Bounds]].
 
Once the variance of the expected reliability is obtained, the two-sided confidence intervals can be calculated using:
 
::<math>[\frac{R}{R+(1-R)w},\frac{R}{R+(1-R)/w}]\,\!</math>
 
where:
:: CL is the confidence level
:: <math>\alpha\,\!</math> is 1-CL
:: <math>w=\exp \{{{z}_{1-\alpha /2}}\sqrt{Var(R)}/[R(1-R)]\}\,\!</math>
:: <math>{{Z}_{1-\alpha /2}}\,\!</math> is the <math>1-\alpha/2\,\!</math> percentile of a standard normal distribution.
 
If the upper bound (U) and lower bound (L) are not infinite and 0, respectively, then the calculated variance of <math> R \,\!</math> is adjusted by <math>{{\left[ {1}/{\left( {{F}_{1}}(U)-{{F}_{1}}(L) \right)}\; \right]}^{2}}\,\!</math>.
 
===Variation in Probability Values===
 
Assume the distributions for stress and strength are known. From the stress-strength equation
 
::<math>R=P[Stress\le Strength]=\int_{0}^{\infty }{{{f}_{Stress}}(x)\cdot {{R}_{Strength}}(x)}dx\,\!</math>
 
we know that the calculated reliability is the '''expected''' value of the probability that a strength value is larger than a stress value. Since both strength and stress are random variables from their distributions, the '''reliability''' is also a random variable. This can be explained using the following example. Let's first assume that stress is a fixed value of 567. The reliability then is:
 
::<math>R(567)=\Pr (Strength>567)={{R}_{2}}(567)\,\!</math>
 
This is the reliability when the stress value is 567 and when the strength distribution is given. If stress is not a fixed value (i.e., it follows a distribution instead), then it can take values other than 567. For instance, if stress takes a value of 700, then we get another reliability value of <math>{R}(700)\,\!</math>. Since stress is a random variable, for any stress value <math>{x}_{i}\,\!</math>, there is a reliability value of
<math>R({{x}_{i}})\,\!</math> calculated from the strength distribution. We will end up with many <math>R({{x}_{i}})\,\!</math>s or <math>R_{2}({{x}_{i}})\,\!</math>s. From these <math>R({{x}_{i}})\,\!</math>s, we can get the mean and variance of the reliability. In fact, its mean is the result from:
 
::<math>R=P[Stress\le Strength]=\int_{0}^{\infty }{{{f}_{Stress}}(x)\cdot {{R}_{Strength}}(x)}dx\,\!</math>
 
and its variance is:
 
::<math>\begin{align}
  & Var\left[ R \right]=Var\left[ {{R}_{2}}({{X}_{1}}) \right]=E\left[ {{R}_{2}}{{\left( {{X}_{1}} \right)}^{2}} \right]-{{\left( E\left[ {{R}_{2}}\left( {{X}_{1}} \right) \right] \right)}^{2}} \\
& =\int_{0}^{\infty }{{{f}_{1}}(x){{\left[ {{R}_{2}}(x) \right]}^{2}}dx}-{{\left( E\left[ {{R}_{2}}\left( {{X}_{1}} \right) \right] \right)}^{2}} \\
& =\int_{0}^{\infty }{{{f}_{1}}(x){{\left[ {{R}_{2}}(x) \right]}^{2}}dx}-{{\left( R \right)}^{2}} \\
\end{align}\,\!</math>
 
where ''R'' is the expected value of the reliability.
 
Once the variance of the expected reliability is obtained, the two-sided confidence intervals can be calculated using:
 
::<math>[\frac{R}{R+(1-R)w},\frac{R}{R+(1-R)/w}]\,\!</math>
 
where:
:: CL is the confidence level
:: <math>\alpha\,\!</math> is 1-CL
:: <math>w=\exp \{{{z}_{1-\alpha /2}}\sqrt{Var(R)}/[R(1-R)]\}\,\!</math>
:: <math>{{Z}_{1-\alpha /2}}\,\!</math> is the <math>1-\alpha/2\,\!</math> percentile of a standard normal distribution
 
 
If the upper bound (U) and lower bound (L) are not infinite and 0, the above calculated variance of ''R'' is adjusted by <math>{{\left[ {1}/{\left( {{F}_{1}}(U)-{{F}_{1}}(L) \right)}\; \right]}^{2}}\,\!</math>.
 
===Example 1===
{{:Stress-Strength Parameter Uncertainty Example}}
 
 
==Stress-Strength Analysis in Design for Reliability==
As we know, the expected reliability is called from the following stress-strength calculation:
 
::<math>R=P[Stress\le Strength]=\int_{0}^{\infty }{{{f}_{Stress}}(x)\cdot {{R}_{Strength}}(x)}dx\,\!</math>
 
The stress distribution is usually estimated from customer usage data, such as the mileage per year of a passenger car or the load distribution for a beam. The strength distribution, on the other hand, is affected by the material used in the component, the geometric dimensions and the manufacturing process.
 
Because the stress distribution can be estimated from customer usage data, we can assume that <math>{f}_{Stress} \,\!</math> is known. Therefore, for a given reliability goal, the strength distribution <math> {R}_{Strength}\,\!</math> is the only unknown in the given equation. The factors that affect the strength distribution can be adjusted to obtain a strength distribution that meets the reliability goal. The following example shows how to use the Target Reliability Parameter Estimator tool in the stress-strength folio to obtain the parameters for a strength distribution that will meet a specified reliability goal.
 
===Example 2===
{{:Stress-Strength Analysis in Design for Reliability}}

Latest revision as of 19:11, 15 September 2023

New format available! This reference is now available in a new format that offers faster page load, improved display for calculations and images, more targeted search and the latest content available as a PDF. As of September 2023, this Reliawiki page will not continue to be updated. Please update all links and bookmarks to the latest reference at help.reliasoft.com/reference/life_data_analysis

Chapter Additional Reliability Analysis Tools: Stress-Strength Analysis


Weibullbox.png

Chapter Additional Reliability Analysis Tools  
Stress-Strength Analysis  

Synthesis-icon.png

Available Software:
Weibull++

Examples icon.png

More Resources:
Weibull++ Examples Collection


Stress-strength analysis has been used in mechanical component design. The probability of failure is based on the probability of stress exceeding strength. The following equation is used to calculate the expected probability of failure, F:

[math]\displaystyle{ F=P[Stress\ge Strength]=\int_{0}^{\infty }{{{f}_{Strength}}(x)\cdot {{R}_{Stress}}(x)}dx\,\! }[/math]

The expected probability of success or the expected reliability, R , is calculated as:

[math]\displaystyle{ R=P[Stress\le Strength]=\int_{0}^{\infty }{{{f}_{Stress}}(x)\cdot {{R}_{Strength}}(x)}dx\,\! }[/math]

The equations above assume that both stress and strength are in the positive domain. For general cases, the expected reliability can be calculated using the following equation:

[math]\displaystyle{ R=P[{{X}_{1}}\le {{X}_{2}}]=\frac{1}{{{F}_{1}}(U)-{{F}_{1}}(L)}\int_{L}^{U}{{{f}_{1}}(x)\cdot {{R}_{2}}(x)}dx\,\! }[/math]

where:

[math]\displaystyle{ L\le {{X}_{1}}\le U\,\! }[/math]
[math]\displaystyle{ \begin{align} & {{X}_{1}}:\text{ Stress } \\ & {{X}_{2}}:\text{ Strength } \\ \end{align}\,\! }[/math]


When U = infinity and L = 0, this equation becomes equal to previous equation (i.e., the equation for the expected reliability R).

Confidence Intervals on the Probability

Both the stress and strength distributions can be estimated from actual data or specified by engineers based on engineering knowledge or existing references. Based on the source of the distribution, there are two types of variation associated with the calculated probability: variation in the model parameters and variation in the probability values. Both are described next.

Variation in Model Parameters

If both the stress and strength distributions are estimated from data sets, then there are uncertainties associated with the estimated distribution parameters. These uncertainties will cause some degree of variation of the probability calculated from the stress-strength analysis. Therefore, we can use these uncertainties to estimate the confidence intervals on the calculated probability. To get the confidence intervals, we first calculate the variance of the reliability based on Taylor expansion by ignoring the 2nd order term. [math]\displaystyle{ R\,\! }[/math], The approximation for the variance is:

[math]\displaystyle{ Var\left[ R \right]=\int_{0}^{\infty }{Var\left[ {{f}_{1}}(x) \right]}{{\left[ {{R}_{2}}(x) \right]}^{2}}dx+{{\int_{0}^{\infty }{\left[ {{f}_{1}}(x) \right]}}^{2}}Var\left[ {{R}_{2}}(x) \right]dx\,\! }[/math]

Variance of [math]\displaystyle{ {{f}_{1}}(x)\,\! }[/math] and [math]\displaystyle{ {{R}_{2}}(x)\,\! }[/math] can be estimated from the Fisher Information Matrix. For details, please see Confidence Bounds.

Once the variance of the expected reliability is obtained, the two-sided confidence intervals can be calculated using:

[math]\displaystyle{ [\frac{R}{R+(1-R)w},\frac{R}{R+(1-R)/w}]\,\! }[/math]

where:

CL is the confidence level
[math]\displaystyle{ \alpha\,\! }[/math] is 1-CL
[math]\displaystyle{ w=\exp \{{{z}_{1-\alpha /2}}\sqrt{Var(R)}/[R(1-R)]\}\,\! }[/math]
[math]\displaystyle{ {{Z}_{1-\alpha /2}}\,\! }[/math] is the [math]\displaystyle{ 1-\alpha/2\,\! }[/math] percentile of a standard normal distribution.

If the upper bound (U) and lower bound (L) are not infinite and 0, respectively, then the calculated variance of [math]\displaystyle{ R \,\! }[/math] is adjusted by [math]\displaystyle{ {{\left[ {1}/{\left( {{F}_{1}}(U)-{{F}_{1}}(L) \right)}\; \right]}^{2}}\,\! }[/math].

Variation in Probability Values

Assume the distributions for stress and strength are known. From the stress-strength equation

[math]\displaystyle{ R=P[Stress\le Strength]=\int_{0}^{\infty }{{{f}_{Stress}}(x)\cdot {{R}_{Strength}}(x)}dx\,\! }[/math]

we know that the calculated reliability is the expected value of the probability that a strength value is larger than a stress value. Since both strength and stress are random variables from their distributions, the reliability is also a random variable. This can be explained using the following example. Let's first assume that stress is a fixed value of 567. The reliability then is:

[math]\displaystyle{ R(567)=\Pr (Strength\gt 567)={{R}_{2}}(567)\,\! }[/math]

This is the reliability when the stress value is 567 and when the strength distribution is given. If stress is not a fixed value (i.e., it follows a distribution instead), then it can take values other than 567. For instance, if stress takes a value of 700, then we get another reliability value of [math]\displaystyle{ {R}(700)\,\! }[/math]. Since stress is a random variable, for any stress value [math]\displaystyle{ {x}_{i}\,\! }[/math], there is a reliability value of [math]\displaystyle{ R({{x}_{i}})\,\! }[/math] calculated from the strength distribution. We will end up with many [math]\displaystyle{ R({{x}_{i}})\,\! }[/math]s or [math]\displaystyle{ R_{2}({{x}_{i}})\,\! }[/math]s. From these [math]\displaystyle{ R({{x}_{i}})\,\! }[/math]s, we can get the mean and variance of the reliability. In fact, its mean is the result from:

[math]\displaystyle{ R=P[Stress\le Strength]=\int_{0}^{\infty }{{{f}_{Stress}}(x)\cdot {{R}_{Strength}}(x)}dx\,\! }[/math]

and its variance is:

[math]\displaystyle{ \begin{align} & Var\left[ R \right]=Var\left[ {{R}_{2}}({{X}_{1}}) \right]=E\left[ {{R}_{2}}{{\left( {{X}_{1}} \right)}^{2}} \right]-{{\left( E\left[ {{R}_{2}}\left( {{X}_{1}} \right) \right] \right)}^{2}} \\ & =\int_{0}^{\infty }{{{f}_{1}}(x){{\left[ {{R}_{2}}(x) \right]}^{2}}dx}-{{\left( E\left[ {{R}_{2}}\left( {{X}_{1}} \right) \right] \right)}^{2}} \\ & =\int_{0}^{\infty }{{{f}_{1}}(x){{\left[ {{R}_{2}}(x) \right]}^{2}}dx}-{{\left( R \right)}^{2}} \\ \end{align}\,\! }[/math]

where R is the expected value of the reliability.

Once the variance of the expected reliability is obtained, the two-sided confidence intervals can be calculated using:

[math]\displaystyle{ [\frac{R}{R+(1-R)w},\frac{R}{R+(1-R)/w}]\,\! }[/math]

where:

CL is the confidence level
[math]\displaystyle{ \alpha\,\! }[/math] is 1-CL
[math]\displaystyle{ w=\exp \{{{z}_{1-\alpha /2}}\sqrt{Var(R)}/[R(1-R)]\}\,\! }[/math]
[math]\displaystyle{ {{Z}_{1-\alpha /2}}\,\! }[/math] is the [math]\displaystyle{ 1-\alpha/2\,\! }[/math] percentile of a standard normal distribution


If the upper bound (U) and lower bound (L) are not infinite and 0, the above calculated variance of R is adjusted by [math]\displaystyle{ {{\left[ {1}/{\left( {{F}_{1}}(U)-{{F}_{1}}(L) \right)}\; \right]}^{2}}\,\! }[/math].

Example 1

Assume that we are going to use stress-strength analysis to estimate the reliability of a component used in a vehicle. The stress is the usage mileage distribution and the strength is the miles-to-failure distribution of the component. The warranty is 1 year or 15,000 miles, whichever is earlier. The goal is to estimate the reliability of the component within the warranty period (1 year/15,000 miles).

The following table gives the data for the mileage distribution per year (stress):

Stress: Usage Mileage Distribution
10096 12405
10469 12527
10955 12536
11183 12595
11391 12657
11486 13777
11534 13862
11919 13971
12105 14032
12141 14138


The following table gives the data for the miles-to-failure distribution (strength):

Strength: Failure Mileage Distribution
13507 16125
13793 16320
13943 16327
14017 16349
14147 16406
14351 16501
14376 16611
14595 16625
14746 16670
14810 16749
14940 16793
14951 16862
15104 16930
15218 16948
15303 17024
15311 17041
15480 17263
15496 17347
15522 17430
15547 17805
15570 17884
15975 18549
16003 18575
16018 18813
16052 18944

Solution

First, estimate the stress and strength distributions using the given data. Enter the stress and strength data into two separate data sheets and analyze each data sheet using the lognormal distribution and MLE analysis method. The parameters of the stress distribution are estimated to be log-mean = 9.411844 and log-std = 0.098741.

Stress-Strength Example 1 Stress-Distribution.png

The parameters of the strength distribution are estimated to be log-mean = 9.681503 and log-std = 0.083494.

Stress-Strength Example 1 Strength-Distribution.png

Next, open the Stress-Strength tool and choose to compare the two data sheets. The following picture shows the pdf curves of the two data sets:

Stress-Strength Example 1 pdf curve.png

Since the warranty is 1 year/15,000 miles, all the vehicles with mileage larger than 15,000 should not be considered in the calculation. To do this, go to the Setup page of the control panel and select the Override auto-calculated limits check box. Set the value of the upper limit to 15,000 as shown next.

Stress-Strength Example 1 Calculation Settings.png

Recalculate the results. The estimated reliability for vehicles less than 15,000 miles per year is 98.84%. The associated confidence bounds are estimated from the variance of the distribution parameters. With larger samples for the stress and strength data, the width of the bounds will be narrower.

Stress-Strength Example 1 Calculation Results.png


Stress-Strength Analysis in Design for Reliability

As we know, the expected reliability is called from the following stress-strength calculation:

[math]\displaystyle{ R=P[Stress\le Strength]=\int_{0}^{\infty }{{{f}_{Stress}}(x)\cdot {{R}_{Strength}}(x)}dx\,\! }[/math]

The stress distribution is usually estimated from customer usage data, such as the mileage per year of a passenger car or the load distribution for a beam. The strength distribution, on the other hand, is affected by the material used in the component, the geometric dimensions and the manufacturing process.

Because the stress distribution can be estimated from customer usage data, we can assume that [math]\displaystyle{ {f}_{Stress} \,\! }[/math] is known. Therefore, for a given reliability goal, the strength distribution [math]\displaystyle{ {R}_{Strength}\,\! }[/math] is the only unknown in the given equation. The factors that affect the strength distribution can be adjusted to obtain a strength distribution that meets the reliability goal. The following example shows how to use the Target Reliability Parameter Estimator tool in the stress-strength folio to obtain the parameters for a strength distribution that will meet a specified reliability goal.

Example 2

Assume that the stress distribution for a component is known to be a Weibull distribution with beta = 3 and eta = 2000. For the current design, the strength distribution is also a Weibull distribution with beta =1.5 and eta=4000. Evaluate the current reliability of the component. If the reliability does not meet the target reliability of 90%, determine what parameters would be required for the strength distribution in order to meet the specified target.

Solution

The following picture shows the stress-strength tool and the calculated reliability of the current design.

Stress-strength example 2 current reliability.png

The result shows that the current reliability is about 74.0543%, which is below the target value of 90%. We need to use the Target Reliability Parameter Estimator to determine the parameters for the strength distribution that, when compared against the stress distribution, would result in the target reliability.

The following picture shows the Target Reliability Parameter Estimator window. In the Strength Parameters area, select eta. Set the Target Reliability to 90% and click Calculate. The calculated eta is 8192.2385 hours.

Stress-strength example 2 Result.png

Click Update to perform the stress-strength analysis again using the altered parameters for the strength distribution. The following plot shows that the calculated reliability is 90%. Therefore, in order to meet the reliability requirement, the component must be redesigned such that the eta parameter of the strength distribution is at least 8192.2385 hours.

Stress-strength example 2 Confirmed Result.png