Falkonry in the US has developed a real time AI tool for anomaly detection in semiconductor manufacturing and power networks using a single Nvidia GPU.
The Falkonry Insight tool operates on high-speed sensor time series data for operations and maintenance teams to make smarter decisions using existing data from machines and processes.
All enterprises, large and small, automate their core processes and operations through a complex orchestration of hundreds to thousands of individual machines and systems. These systems generate terabytes of time series data containing signs of deteriorating machine health and product quality.
Falkonry Insight allows operations teams to virtually inspect every electrical and mechanical data point to automatically detect new and emerging hotspots.
The sensor-based deep learning AI system was developed with Nvidia using an optimized TensorRT-based inference runtime capable of exploiting the full capacity of its GPU processors. A factory-scale edge-based system which might have required dozens of dual-socket servers to support a CPU-based inference pipeline can now be provisioned with a modest three-node cluster incorporating a single NVIDIA A10 Tensor Core GPU. This can process terabytes of data at plant scale to surface machine and process anomalies without the need for any manual intervention.
Falkonry Insight organizes detected anomalies by the affected components of the plant, the severity of the anomaly, and the main contributing factors. It also provides a collaborative reporting environment where behaviours can be confirmed and catalogued by operations and maintenance teams using rich time series visualization capabilities.
“Autonomous industrial AI systems that can be directly used and managed by operations and maintenance teams are essential for realizing the benefits of smart manufacturing. From years of experience working with smart factories, Falkonry has been able to design such a system to dramatically change the rate of adoption across a spectrum of industries,” said Dr. Nikunj Mehta, Founder and CEO, Falkonry. “We could not have achieved this level of scale and manageability of time series AI without our collaboration with NVIDIA and are excited to continue working closely with the company.”
“Real-time AI is a critical element of the industrial metaverse, and accelerated computing is essential to surfacing the live state of physical systems from massively underutilized quantities of time series data on the shop floor,” said Piyush Modi, Industrial Business Development Lead at NVIDIA. “Falkonry’s time series AI is well positioned to benefit from NVIDIA’s investment in deep learning at the scale of entire plants and enabling highly interactive, live digital twins in NVIDIA Omniverse Enterprise for managing complex industrial operations.”
Falkonry was founded in 2012 and has 12 patents across the world for its advanced pattern recognition technology. Falkonry Insight will be available to customers in Semiconductor Manufacturing and Power Generation starting November 1st, 2022 as well as Oil & Gas, Automotive, Pharmaceuticals and Chemicals.
The technical storage or access is strictly necessary for the legitimate purpose of enabling the use of a specific service explicitly requested by the subscriber or user, or for the sole purpose of carrying out the transmission of a communication over an electronic communications network.
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user.
The technical storage or access that is used exclusively for statistical purposes.The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes.