Condition based monitoring of turbomachinery components
Transcript of Condition based monitoring of turbomachinery components
![Page 1: Condition based monitoring of turbomachinery components](https://reader033.fdocuments.in/reader033/viewer/2022041710/62538adbaeaecd76272de268/html5/thumbnails/1.jpg)
“TURBO-REFLEX. TURBOmachinery REtrofit enabling FLEXible back-up capacity for the transition
of the European energy system”
Condition based monitoring of turbomachinery components
Alexander Wiedermann, MAN Energy Solutions
ETN Webinar Series
FLEXIBLE POWER GENERATION
March 2nd, 2021, 12am-1pm CET
![Page 2: Condition based monitoring of turbomachinery components](https://reader033.fdocuments.in/reader033/viewer/2022041710/62538adbaeaecd76272de268/html5/thumbnails/2.jpg)
2
Tasks
ETN Webinar Series FLEXIBLE POWER GENERATION March 2nd, 2021, 12am-1pm CET
4.1 Condition and efficiency monitoring system
− Monitoring system based on analytics and test data
− Detection of critical operating conditions
− Prediction of maintenance intervals
4.2 Steam turbine monitoring system
− Monitoring system for blade vibrations
− Monitoring and analysis tool for LCF
− Online rub monitoring system
4.3 Power generation analytics
− Probes for plant components
− Plant monitoring system
4.4 Machine learning on large heterogeneous
data
− Big data analytics methods
− Employ machine learning algorithms
Task 4.3
Task 4.2 Task 4.1; Task 4.4
![Page 3: Condition based monitoring of turbomachinery components](https://reader033.fdocuments.in/reader033/viewer/2022041710/62538adbaeaecd76272de268/html5/thumbnails/3.jpg)
Condition based monitoring: Vision
ETN Webinar Series FLEXIBLE POWER GENERATION March 2nd, 2021, 12am-1pm CET3
![Page 4: Condition based monitoring of turbomachinery components](https://reader033.fdocuments.in/reader033/viewer/2022041710/62538adbaeaecd76272de268/html5/thumbnails/4.jpg)
4
CBM and ML: Approach
ETN Webinar Series FLEXIBLE POWER GENERATION March 2nd, 2021, 12am-1pm CET
“Machine learning on large heterogeneous data sources”
to develop and use plant physics models to optimize flexibility improvement strategies
to prepare, implement and validate problem specific machine learning algorithms and feature
tailored methods
Synopsis
PCA: Principal Component Analysis
STFT: Short Time Fourier Transform
CWT: Continuous Wavelet Transform
SVM: Support Vector Machine
![Page 5: Condition based monitoring of turbomachinery components](https://reader033.fdocuments.in/reader033/viewer/2022041710/62538adbaeaecd76272de268/html5/thumbnails/5.jpg)
Example: GT Start-up failures
ETN Webinar Series FLEXIBLE POWER GENERATION March 2nd, 2021, 12am-1pm CET5
Predictive capability of a trained CNN- Model (Convolutional Neural Network):
Probability figures show the ratio of true vs. false predictions for each difference
image
![Page 6: Condition based monitoring of turbomachinery components](https://reader033.fdocuments.in/reader033/viewer/2022041710/62538adbaeaecd76272de268/html5/thumbnails/6.jpg)
An OEM Consortium of
25 partners in 9 countries
6
Contact person:
Alexander Wiedermann
Phone:
00 49 208 692 2238
Fax:
00 49 208 692 2958
Email:
ETN Webinar Series FLEXIBLE POWER GENERATION March 2nd, 2021, 12am-1pm CET