Falkonry

Plant operators use the time series AI of Falkonry to watch sensor and machine data across their operations and to find events that would otherwise stay hidden. The company says its platform looks at many variables at once, including those spread across different systems and product changes, and it aims to show a live, accurate view of how a plant is running.

The company argues that its platform can identify events that an operations team cannot see, explain why they happened, and drive action to fix them. It says the result is higher availability, fewer defects, a more engaged workforce and better performance. Its stated aim is to make operational reliability a matter of evidence, so that decisions can be taken on the basis of what the data shows.

Predictive maintenance and condition monitoring

Predictive maintenance is the company's most prominent use case. The company says its platform raises alerts at the earliest sign of a developing fault, which lets an operator plan repairs instead of reacting to a breakdown. Its examples include predictive maintenance in vertical casting and rolling operations, condition based maintenance for mission reliability, and continuous monitoring of equipment in semiconductor fabrication. One example says compressor seal failures can be predicted weeks in advance.

Quality assessment and yield improvement

Quality assessment and yield improvement cover the detection of problems that lead to scrap or defects. The company gives an example of weld quality assessed in real time, which it says reduces the number of faulty welds that escape the line. Yield work is described as finding the causes of losses in production and presenting them so that the team can act on them.

Root cause analysis from machine and process data

The company says its explanations help operators uncover the hidden causes behind production losses. Operations teams are often short of the time needed to analyze and diagnose problems, and the company presents its platform as a way to speed up root cause analysis. Its approach is to explain patterns in the data, rather than only flagging that something has changed.

Capturing expert knowledge in the platform

The platform is described as able to capture expert knowledge and recall it later, so that a plant's know how is not lost when people move on. The company says this is one step towards greater autonomy in operations, where the system can recognize patterns the experts have already identified.

Data sources and industrial connections

The company describes ingestion from programmable logic controllers, industrial buses, SCADA and distributed control systems, and from physics sensors that measure power quality. The platform is presented as handling high speed streams of data from these sources and turning them into multivariate patterns across the assets in a plant. The company says its sectors include energy production, factories, buildings, vehicles and software operations.

Operational resilience for mission critical systems

The company describes its platform as an operational resilience product for systems that cannot afford unplanned downtime. It presents the aim of moving from raw sensor data to fewer unplanned stoppages, and it describes its work as useful wherever a plant's assets have to keep running safely and reliably.