Structural health monitoring and ice detection for connected wind turbines with Wendelin
Dr.-Ing. Steffen Pankoke
Pankoke (at) woelfel (dot) de
Dr. Klaus Wölfel
klaus (at) nexedi (dot) com
Structural health monitoring and ice detection for connected wind turbines with Wendelin
Dr.-Ing. Steffen Pankoke
Pankoke (at) woelfel (dot) de
Dr. Klaus Wölfel
klaus (at) nexedi (dot) com
IT Problems
- Reliable data collection
- Scalable storage
- Parallel processing
- Out-of-core processing
- Predictive algorithms
Introducing Wendelin
Reliable data collection: Fluentd
Scalable storage: NEO
Parallel Processing: ERP5
# Initialize data
data_size = 1e6
server_count = 1000
chunk_size = data_size / server_count
data = zbigarray(data_size)
# Process data in parallel on each server (Map Reduce, Batch, etc)
for server in server_count:
data.activate().process(server*chunk_size, chunk_size)
Out-of-core processing: wendelin.core
# Numpy
np.ndarray(shape=(2,2), dtype="float")
# Out-of-core data
ZBigArray(shape=(1e18,2), dtype="float")
# Full out-of-core
ZBigArray(shape=(1e9,2e9), dtype="float")
Predictive algorithms: scikit-learn
User Interface
Data Visualisation
Thank You
- Nexedi GmbH
- Agnes-Pockels-Bogen 1
- 80992 München
- Germany