Method for training a machine learning model usable for determining a remaining useful life of a wind turbine
The application relates to a method, in particular a computer-implemented method, for training a machine learning model usable for determining a remaining useful life of a wind turbine, including providing a plurality of operation data sets of a reference wind turbine, providing a plurality of load data sets of the reference wind turbine, wherein a load data set is based on at least one load parameter measured at the reference wind turbine, and generating a plurality of wind turbine training data sets for training a machine learning model by synchronously assigning a respective operation data set with a respective load data set.
1 . A computer-implemented method for training a machine learning model used for determining a remaining useful life of a further wind turbine, comprising:
providing a plurality of operation data sets of a reference wind turbine;
providing a plurality of load data sets of the reference wind turbine, wherein a load data set is based on at least one load parameter measured at the reference wind turbine; and
generating a plurality of wind turbine training data sets for training the machine learning model by synchronously assigning a respective operation data set with a respective load data set,
wherein the plurality of wind turbine training data sets are in the form of text data, from which a presentation in form of a vector, a matrix, and/or a tensor is generated, and
wherein the step of generating the plurality of wind turbine training data sets comprises the step of generating approximately 100.000 wind power training data sets; and
training the machine learning model on the plurality of wind turbine training data sets.
2 . The method according to claim 1 , wherein the step of training the machine learning model comprises the step of training the machine learning model on the plurality of wind turbine training data sets during a training time period.
3 . The method according to claim 2 , wherein
at least one regularization technique is applied during the training time period,
wherein the regularization technique is selected from a group comprising:
linear Bayesian regression,
Bayesian neural network with concrete dropout,
Bayesian neural network with variational inference
adaptive Bayesian spline regression.
4 . The method according to claim 2 , wherein
a portion of the generated wind turbine training data sets is used as validation data sets during the training time period.
5 . The method according to claim 4 , wherein
during the training time period, a model for estimating aleatory and/or epistemic uncertainties is applied, wherein the estimation is achieved by applying at least one machine learning method, wherein the machine learning method is selected from a group comprising:
linear Bayesian regression,
Bayesian neural network with concrete dropout,
Bayesian neural network with variational inference
adaptive Bayesian Spline Regression, and/or
wherein the estimation is implemented by bootstrapping.
6 . The method according to claim 1 , wherein
a provided operation data set contains at least one operation parameter of the reference wind turbine,
wherein the at least one operation parameter is selected from a group comprising:
tower top acceleration,
pitch angle,
pitch speed,
rotor speed,
electrical power,
nacelle wind speed.
7 . The method according to claim 6 , wherein
an operation data set comprises as operation parameter value of an operation parameter at least one operation parameter value selected from a group comprising:
maximum operation parameter value detected during an operating data time period,
minimum operation parameter value detected during the operation data time period,
operation parameter mean value determined from the operation parameter values detected during the operation data time period,
standard deviation determined from the operation parameter values detected during the operation data time period.
8 . The method according to claim 1 , wherein
a provided load data set is based on at least one measured load parameter of the reference wind turbine,
wherein the load parameter is selected from a group comprising:
blade root load parameter,
rotor load parameter,
tower load parameter,
tower torsion parameter,
tower top moment parameter.
9 . The method according to claim 8 , wherein the method comprises:
measuring the at least one load parameter of the reference wind turbine.
10 . The method according to claim 9 , wherein
measuring the at least one load parameter is performed according to the IEC 61400-13 standard [2], Edition 1, 2001.
11 . The method according to claim 8 , wherein
as load parameter value of a load parameter a load parameter value is provided, selected from a group comprising:
maximum load parameter value measured during a load data time period,
minimum operation parameter value measured during the load data time period,
load parameter mean value determined from the load parameter values measured during the load data time period,
standard deviation determined from the load parameter values measured during the load data time period.
12 . The method according to claim 1 , wherein the method comprises:
measuring the at least one load parameter of the reference wind turbine during a measurement time period; and
detecting operational data parameter values of the reference wind turbine during the measurement time period.
13 . The method according to claim 12 , wherein
the measurement time period is at least 3 months.
14 . The method according to claim 1 , wherein the method comprises:
forming a load data set by converting the at least one measured load parameter value of the reference wind turbine into at least one fatigue load indicator.
15 . The method according to claim 14 , wherein
the converting is based on a Rainflow Counting method.
16 . The method according to claim 1 , wherein
the machine learning model is and/or comprises an artificial neural network.
17 . A method of using a machine learning model trained according to claim 1 , comprising:
inputting at least one operation data set of the further wind turbine into the machine learning model, and
outputting, by the machine learning model, at least one turbine condition data set.
18 . The method according to claim 17 , wherein
the further wind turbine is a wind turbine type identical to the wind turbine type of the reference wind turbine,
and/or
the further wind turbine and the reference wind turbine are comprised by the same wind farm.
19 . The method according to claim 17 , wherein the method comprises:
determining the remaining runtime of the further wind turbine based on the at least one turbine condition data set of the further wind turbine.
20 . A computing device comprising at least one data memory containing computer program code and at least one processor, wherein the data memory and the processor are configured such that the computing device is caused to generate at least one turbine condition data set based on at least one provided operation data set and at least in part using a machine learning model trained according to claim 1 .