Data-based Reliability Analysis And Failure Rate Prediction Of Wind Farm Turbine Gearboxes

Gearbox is one of the most important components of wind farm turbine.Gearboxes tend to have a higher failure rate and longer downtime due to the operating time and status.Research on the reliability analysis of gearboxes will provide a certain reference for wind farms to formulate gearbox inventory management and maintenance cost plans.In traditional industrial reliability models,the failure rate is generally studied over time.While this thesis introduces the state of wind turbine on the basis of taking time as the independent variable,and further popularizes the traditional reliability model to study the reliability changes of the gearbox during operation time and state.When more actual data is collected,the model parameters can be updated and revised to achieve datadriven.Firstly,this thesis analyzes the constant failure rate model.Analyze the reliability function,the probability density function and the cumulative distribution function under the model.By setting different failure rates and calculating the failure probability of the gearbox,it can be found that when the failure rate is constant,the failure probability curve of the gearbox increases monotonously;at a certain time,the failure probability of the gearbox increases with the increase of the failure rate.Numerical simulation experiments are performed by assuming that the fault distribution of all gearboxes in the wind farm is independent and identical.A binomial distribution model is established,the probability of a specific number of wind farms failing at a specific time is calculated,and the corresponding curve is drawn.The curve shape conforms to the binomial distribution nature.Moreover,the gearbox failure probability and the mathematical expectation of failures is calculated every year.Drawn the curve,and found that the curve monotonously rises with time.Secondly,this thesis promotes the above-mentioned models,studies the situation that the failure rate of wind farm gearboxes is constant during the life cycle but not exactly the same as each other.The models under two failure rates are studied,through numerical simulation experiments,the binomial distribution model is used to calculate the probability of a specified number of failures at a specified time,and the curve is drawn.It is found that the curve also presents the characteristics of a binomial distribution curve,that is,it first rises monotonically,and then drops after reaching the peak.And as time increases,the peak of the curve gradually shifts to the right.By calculating the mathematical expectation of failure numbers per year,it is also found that as time increases,the failure numbers gradually increases,the ascent rate gradually slows down.Then,the failure rate numbers is extended to n,derived the probability of a specified number of gearbox failures at a specified time when the failure rate of the gearbox is constant but different from each other.Through numerical simulation,the average value of gearboxes failure probability and the mathematical expectation of failure numbers are calculated.Both curves rise monotonously,and the rate of rise gradually slows down.Further,this thesis studies the reliability model when the failure rate is not constant,but changes with time.Analyzes the Weibull distribution and assumed that each gearbox is independent and identically,and the failure numbers obeys the binomial distribution.Draw the failure rate curve by numerical simulation in the shape of a "bathtub curve".The relationship curve between the failure numbers and the probability also shows a trend of first rising and then falling,with the peak value shifting to the right over time.This thesis uses Sichuan wind farm maintenance data for case analysis.The data contains three wind farm data.The failure rate was calculated using the time to failure,and the Weibull failure distribution model was respectively fitted,and the probability of different failure numbers per year was calculated and predicted.Draw the curves and calculate the mathematical expectation.By comparing the calculation results,the relationship between the reliability of the three wind fields is obtained.The results can provide a certain reference for the practical application.Finally,this thesis continues to promote the failure rate model.When the failure rate obeys the Weibull distribution,the gearbox state are added to study the impact on the failure rate.Using the combination of number and shape and numerical analysis,from the perspective of how the status affects the failure rate "bath curve",four reliability models are constructed,and the reliability model failure rate,reliability function,and probability of failure distribution are derived.According to the density function formula,and cumulative distribution function,it is concluded that the four state can respectively represent the influence on the overall failure rate,the overall failure period time,the failure rate during the failure period,and the lifetime of the failure period.Numerical simulations were carried out to study the influence of gearbox status on the number of failures and failure time under different occasions.The results show that the worse the gearbox status is,the more the relationship curve between the failure numbers and the probability shifts to the right,which tends to have more failures during the failure period.The results of the model have certain rationality and practical significance.It is recommended that the wind farm pay attention to the status of the gearbox and prepare in advance to reduce losses.