Siemens Energy automates detection of leaks and abnormal noises in power plants with AI
Siemens Energy
AWS
Safety Inspections
NVIDIA GPUs on AWSs, NVIDIA Triton Inference Server
Siemens Energy is leveraging AI to deliver a path to autonomous power plants, reducing costs along the way. Hundreds of inspection types are usually performed by human walk-throughs, requiring domain expertise. Additionally, many power plants are no longer in always-on mode and, therefore, don’t require 24/7 staffing. As a result, Siemens is developing autonomous power plants, starting with automated remote monitoring for liquid, steam, and oil leaks, as well as abnormal noises.
Siemens Energy runs multiple deep learning models using NVIDIA Triton™ Inference Server on Amazon Web Services (AWS). These models process streaming data from cameras and sensors at the power plants for real-time inference of various safety inspections. Model ensembles enabled by Triton Inference Server also allow for additional pre-processing of images, such as person anonymization.
Results
NVIDIA Triton Inference Server enables standardized deployment across AI frameworks and different models. The solution has proven capable of handling millions of sensors and massive input streams. Triton also makes it easy to deploy on the cloud or at the edge, which is beneficial when data can’t be moved out of the power plants or when edge analytics are required. Now Siemens can automate hundreds of inspection types.
Siemens Energy is meeting the growing global demand for energy while transitioning to a more sustainable world. They offer solutions across the entire energy value chain, including generation, transmission, industrial applications, renewable energy, and more.
“The flexibility of NVIDIA Triton Inference Server is enabling highly complicated power plants, often equipped with cameras and sensors but with legacy software systems, to join the autonomous industrial revolution taking place.”
Product Manager
Siemens Energy
