Reliability-based lifetime fatigue damage assessment of offshore composite wind turbine blades

arXiv:2404.10021 · eess.SY, cs.SY · Submitted 2024-04-15 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Robotics Radio. Generated commentary on the latest robotics and control papers.

Rosa: I'm Rosa.

Dev: And I'm Dev, here with Taro.

Rosa: We're looking at the paper: "Reliability-based lifetime fatigue damage assessment of offshore composite wind turbine blades".

Dev: Here's the gist of the paper. Monitoring Based Fatigue Damage Prognosis of Wind Turbine Composite Blades under Uncertain Wind Loads presents a methodology for analyzing fatigue failure probability in composite blades by using monitoring stochastic deterioration modelling The results show that the stochastic fatigue damage model can give reliable results for time-dependent reliability analysis for composite blades of wind turbines.

Rosa: Lifecycle assessment of wind turbines is essential to improve their design and to optimum maintenance plans for preventing failures during the design life A critical element of wind turbines is the composite blade due to uncertain cyclic wind loads with relatively high frequency and amplitude in offshore environments It is critical to detect the wind fatigue damage evolution in composite blades before they fail catastrophically and destroy the entire wind turbines.

Dev: The challenge in fatigue damage evolution is that the fatigue damage cannot be accurately predicted as deterministic values because of various uncertainties in the harsh offshore environments Therefore, it is vital to select a structural health monitoring system to assess the fatigue damage behaviour of wind turbine blades.

Rosa: Modeling Wind Loads and Internal Stresses The analysis of wind load on the composite blades is needed to ensure whether the turbine resists the action of wind fatigue load within an appropriate safety range The aerodynamic load can be simplified as wind pressure P on one side of composite blade.

Paper summary: Rosa: Looking at the full picture of "Reliability-based lifetime fatigue damage assessment of offshore composite wind turbine blades", it really comes down to using this stochastic gamma process model as a tool for time-dependent reliability analysis of composite blades. The authors show that this approach is effective in predicting the failure probability under varying conditions.

Dev: They establish that the stochastic fatigue damage model can give reliable results for time-dependent reliability analysis for composite blades, and they tested this with a numerical example of a wind turbine composite blade to show its applicability in practice.

Taro: So what we have here is a method where you take real measurements of wind speed and use finite element analysis to get stress values, then plug those into this damage model to see the failure probability over time.

Rosa: Right, and what this changes for us is that we can set maintenance requirements based on these calculated probabilities. The paper's focus is on extending the service life of composite blades under proper maintenance by reducing the risk of structural failure.

Dev: Essentially, they are giving engineers a way to quantify the risk and decide when to intervene, which is crucial because if you don't maintain it at that point, there's a very high probability of catastrophic failure.

Conclusion: Rosa: So we've been looking at how researchers are trying to figure out if these massive wind turbine blades will actually last their whole designed life without failing catastrophically.

Dev: Yeah, this paper, "Reliability-based lifetime fatigue damage assessment of offshore composite wind turbine blades," is basically about using a smart way to predict that failure chance.

Rosa: It uses this method called the stochastic gamma process model to track how damage builds up over time based on real wind loads.

Dev: They show that this model gives reliable results for predicting the failure probability, even when you have all those uncertain wind conditions out there offshore.

Rosa: The main point is that you can use this to set maintenance plans because if the probability of failure gets too high, you know it's time to fix something before things go wrong.

Dev: It means engineers can move away from just guessing and start having a data-driven way of deciding when to inspect or repair those blades.

Rosa: It’s about taking that huge uncertainty in offshore wind and turning it into a manageable risk number you can act on.

Dev: It connects the material science of the composite blade directly to the operational needs of keeping that turbine spinning safely for years.

Rosa: So, this work gives us a solid way to look at those blades and understand their real lifespan under constant stress.

Dev: And it opens up a whole new way to build safety cases for these structures in the real world.

Department of Engineering & Science, University of Greenwich

eess.SY, cs.SY

Submitted: 2024-04-15

Updated: 2026-11-03

Journal ref: Zhang, C., Chen, H. P., Tee, K. F., & Liang, D. (2021). Reliability-based lifetime fatigue damage assessment of offshore composite wind turbine blades. Journal of Aerospace Engineering, 34(3), 04021019

DOI: 10.1061/(ASCE)AS.1943-5525.000126

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 74/100

The gist: Monitoring Based Fatigue Damage Prognosis of Wind Turbine Composite Blades under Uncertain Wind Loads presents a methodology for analyzing fatigue failure probability in composite blades by using…

Key concepts

Stochastic Fatigue Damage Model
This is a mathematical model that uses randomness (stochastic processes) to predict how damage accumulates in composite blades over time. Instead of predicting damage as a fixed number, it accounts for the uncertainties in wind loads, making the prediction more realistic for real-world conditions.
Gamma Process
The gamma process is a specific type of continuous stochastic process used to model gradual damage like fatigue or corrosion. It has properties that make it suitable for modeling how damage slowly builds up over time in materials subjected to repeated stresses, offering straightforward and trustworthy mathematical calculations.
Fatigue Damage Evolution
This describes the non-linear way composite materials degrade under cyclic wind loads. The damage isn't constant; it changes based on the stress cycles experienced. The model tracks this evolution to determine when the accumulated damage reaches a critical level that could lead to catastrophic failure.
Reliability-Based Lifetime Assessment
This approach uses probability to assess how long a structure, like a wind turbine blade, is expected to function safely. By calculating the probability of failure over time based on the damage model, engineers can determine when maintenance is critically needed to prevent structural failure.

Terminology

Summary

Monitoring Based Fatigue Damage Prognosis of Wind Turbine Composite Blades under Uncertain Wind Loads presents a methodology for analyzing fatigue failure probability in composite blades by using monitoring stochastic deterioration modelling The results show that the stochastic fatigue damage model can give reliable results for time-dependent reliability analysis for composite blades of wind turbines. p. 0

Introduction to the Problem

Lifecycle assessment of wind turbines is essential to improve their design and to optimum maintenance plans for preventing failures during the design life A critical element of wind turbines is the composite blade due to uncertain cyclic wind loads with relatively high frequency and amplitude in offshore environments It is critical to detect the wind fatigue damage evolution in composite blades before they fail catastrophically and destroy the entire wind turbines. The challenge in fatigue damage evolution is that the fatigue damage cannot be accurately predicted as deterministic values because of various uncertainties in the harsh offshore environments Therefore, it is vital to select a structural health monitoring system to assess the fatigue damage behaviour of wind turbine blades. p. 1

Modeling Wind Loads and Internal Stresses

The analysis of wind load on the composite blades is needed to ensure whether the turbine resists the action of wind fatigue load within an appropriate safety range The aerodynamic load can be simplified as wind pressure P on one side of composite blade. The values for air density and pressure coefficient are taken as 1. p. 229 kg/m3 and 2.0, respectively. p. 2 p. 2 The maximum internal stress in composite blades can be found in element 787 by ABAQUS 6. p. 314.

Fatigue Damage Evolution Model

The fatigue damage evolution is non-linear in composite materials and the development of fatigue process is shown in Figure 3 The damage equation of model [8] can be written as D(n) = 1 − (1 − (n/N) A)B. The relationship between A and B can be simply written as A = p + q. p. 4 The fatigue life cycles are usually from the S–N curve according to the experimental data. p. 4

Failure Probability Prediction

The relationship between the fatigue damage and the cycles of stress given in Equation 2 can be used for composite blades. The probability density function can be written as (D) = G(Dv, u) = u v / Γ(v) v−1e−∫ (0,∞)(D). p. 5 The equation for the failure probability can be calculated from [4, 11]. p. 5

Numerical Example and Conclusions

A numerical case study is presented to investigate the effectiveness of this stochastic method The results show that stochastic fatigue damage model gives reliable results and can be used for analysing the failure probabilities of composite blades. The failure probability under different defined critical fatigue can help evaluate the failure time of composite blades of wind turbines for inspection and maintenance in various situations. p. 7 p. 7

Acknowledge

The authors would like to thank Dr Weifei Hu of Cornell University for providing the composite blade of finite element model for this study.

References

[1] Jureczko, M.E.Z.Y.K., Pawlak, M. p. 7 and Mężyk, A. p. 7, 2005

[2] Mallick, P.K., 2007

[3] Blanco, M.I., 2009

[4] Van Noortwijk, J.M., 2009

[5] Guida, M. and Penta, F. p. 7, 2015

[6] Hu, W., Han, I., Park, S.C. and Choi, D. p. 7H., 2012

[7] Hu, W., Choi, K.K., Zhupanska, O. p. 7 and Buchholz, J. p. 7H., 2015

[8] Reifsnider, K.L., Henneke, E. p. 7G., Stinchcomb, W. p. 7W. and Duke, J. p. 7C., 1983

[9] Wu, F. and Yao, W. p. 4, 2010

[10] Fuqiang, W. and Weixing, Y. p. 7, 2008

[11] Chen, H.P. and Alani, A. p. 7M., 2013

--- Page 8 ---

The probability of failure increases dramatically over time and reaches approximately 90% at the time when the expected fatigue damage exceeds the given acceptable limit. p. 6 The requirement for maintenance becomes critical to reduce the risk of structural failure and to prevent the unacceptable possible loss before reaching this stage. p. 5 p. 5 Thus the service time of composite blades can be extended under proper maintenance. p. 5

--- Page 7 ---

The trend of the probability of failure for four lines are similar. With increasing the service time, the failure probability increases gradually until reaching certain time points, and then the curve has a rapid rise as the fatigue damage has reached the defined critical damage. p. 6 Finally, the failure probability reaches to a value of very close to unity where the structures need to be maintained. p. 6

--- Page 4 ---

The maximum internal stress in composite blades can be found in element 787 by ABAQUS 6.14. The values for wind amplitude stress by Equation 1 on composite blades as shown in Figure 1. p. 2

--- Page 3 ---

There are three components in the blades finite element model, i.e. rotor, skin and shear web. p. 2 The shear web can prevent the shear forces and have a positive influence on the resistance of composite blades for bending. p. 2

--- Page 5 ---

Stiffness degradation fatigue models for composite laminates have been widely investigated theoretically and experimentally investigations, and they can predict the damage process in the initial or/and middle period of the fatigue life. However, they are not capable of fitting the damage progress in the whole period, as shown in Figure 3. p. 3

--- Page 1 ---

The lifecycle assessment of wind turbines is essential to improve their design and to optimum maintenance plans for preventing failures during the design life. A critical element of wind turbines is the composite blade due to uncertain cyclic wind loads with relatively high frequency and amplitude in offshore environments. p. 0

--- Page 2 ---

In order to improve the performance of wind turbines blades, layered fibre-reinforced polymer composite materials are usually manufactured for the large blades of wind turbines since these materials have better fatigue resistance and lighter weight than traditional materials, e.g., metals. It is important to detect the blade fatigue damage before the blades fail catastrophically, which may destroy the entire wind turbine. p. 1

--- Page 4 ---

For simplicity, it is considered that the aerodynamic load can be simplified as wind pressure P on one side of composite blade. The values ρ and Cp can be taken as 1. p. 229 kg/m3 and 2.0, respectively.

--- Page 5 ---

The maximum stress from finite element model is 718MPa and the ultimate stress is 1548MPa.

--- Page 6 ---

The fatigue damage in composite contains both microscopic and macroscopic mechanisms at all stages during the fatigue process. According to Reifsnider [8], the fatigue damage evolution is non-linear in composite materials and the development of fatigue process is shown in Figure 3. p. 3

--- Page 7 ---

The probability density function Ga(xv,u) is given by G(v, u) = uv / Γ(v) v−1e−∫ (0,∞). The complete gamma function Γ(v) (v≥0) and incomplete gamma function Γ(v,x) (v≥0 and x>0) is defined as. p. 5

--- Page 8 ---

The scale parameter u could be estimated from statistical estimation methods such as a maximum likelihood method and method of moments. The probability of failure per unit time at ti is computed from p i = F(t i) − F(t i−1), f i = 1,2 … T. p. 5

--- Page 7 ---

The equation for the failure probability can be calculated from [4, 11]. The fatigue failure is defined as experiencing n times loading at t n time and when the composite suffers N times, the fatigue damage reaches the unit damage. p. 5

--- Page 8 ---

The service life of the wind turbine blades can be predicted by accumulating the increase growth of fatigue damage in each time before reaching assumed critical fatigue damage. The maintenance for structural repairing therefore should be undertaken in time to prevent structural failure. p. 5

--- Page 8 ---

The fatigue failure probability of the structure increases as the resistance for blades reduces. The probability of failure reaches unity and the structure fails. p. 5

--- Page 8 ---

The fatigue damage process is a continuous stochastic process with the following three properties (1) X(t)=0 with probability one; (2) X(t) has independent increments; (3) X(t)-X(s) Ga(v(t-s),u) for all t>s≥0.

--- Page 8 ---

The gamma process is a continuous stochastic process with the following three properties (1) X(t)=0 with probability one; (2) X(t) has independent increments; (3) X(t)-X(s) Ga(v(t-s),u) for all t>s≥0.

--- Page 8 ---

The gamma process is suitable to model gradual damage such as wear, fatigue, corrosion, erosion. The advantage of this stochastic process is that the required mathematical calculations are relatively straightforward and the results are trustful. p. 4

--- Page 8 ---

The service time of composite blades can be extended under proper maintenance. The requirement for maintenance becomes critical to reduce the risk of structural failure and to prevent the unacceptable possible loss before reaching this stage.

--- Page 8 ---

The probability of failure associated with the fatigue damage depends on the given acceptable limit, with a higher probability of failure for a lower acceptable level at any given time. The probability of failure increases dramatically over time and reaches approximately 90% at the time when the expected fatigue damage exceeds the given acceptable limit.

--- Page 8 ---

The stochastic gamma process model gives good simulations on fatigue damage evolution of wind turbine composite blades. The proposed stochastic fatigue damage model based on the gamma process can be used for time-dependent reliability analysis. p. 7 The failure probability under different defined critical fatigue can help evaluate the failure time of composite blades of wind turbines for inspection and maintenance in various situations.

--- Page 8 ---

The results show that stochastic fatigue damage model can give reliable results for time-dependent reliability analysis for composite blades. The numerical example of a wind turbine composite blade is investigated to show the applicability of the proposed model. p. 0 p. 0

--- Page 8 ---

The results show that the stochastic fatigue damage model can give reliable results for time-dependent reliability analysis for composite blades. The numerical example of a wind turbine composite blade is investigated to show the applicability of the proposed model.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improve predictive maintenance scheduling by using a stochastic gamma process model to predict failure probabilities over time, as described in equation 9: P(t >= tN) = Integral (D(t)) dt. This allows for proactive maintenance planning based on the time-dependent reliability analysis mentioned in the paper.

  2. Develop a structural health monitoring system that integrates finite element analysis results with wind load uncertainties to accurately obtain internal stresses, as suggested by the text: On the basis of 5 minutes mean wind speed measurements, the internal stresses can be accurately obtained from finite element analysis. This system will provide real-time damage assessment for composite blades.

  3. Enhance fatigue life prediction capabilities by incorporating a stiffness degradation rule derived from Reifsnider's model, D(n) = 1 - (1 - (n/N) A)B, into the stochastic framework. This improves the model's ability to fit the damage progress in the whole period of fatigue life rather than just initial stages.

  4. Create a system that estimates required maintenance intervals by analyzing failure probabilities under different critical damage levels, such as Dcr=70%, 80%, 90% and 95%. This directly addresses the conclusion: The failure probability under different defined critical fatigue can help evaluate the failure time of composite blades.

  5. Implement a method to estimate the scale parameter u of the gamma process using statistical estimation methods like maximum likelihood or method of moments. This will make the stochastic model more robust by providing a data-driven way to quantify uncertainties in real operation environments.

Related papers