Please note that JavaScript and style sheet are used in this website,
Due to unadaptability of the style sheet with the browser used in your computer, pages may not look as original.
Even in such a case, however, the contents can be used safely.
Complex Systems Analytics
We are seeing a number of manufacturing plants, power plants and any other complicated and large systems being complicated and complicated recent years. Due to this new situation happening on those systems where highly separated operations, black-boxed component and further complex interactive behaviors are observed, it is getting harder and harder even for expertise on those systems to understand the whole behavior of the systems.
With these situations in mind, NEC has been working on developing new research technologies to profile system characteristics by utilizing massive sensor data collected from many components of the systems. One of the technologies we have developed is System Invariant Analysis Technology, called SIAT, which takes a very unique approach to understand complicated system behaviors. SIAT automatically extracts every local dependency among collected sensor data and accumulate them to describe the system behaviors.
SIAT automatically discovers invariants; a set of constant relationships between two time series of sensor data, without any completed configuration/parameter settings and domain specific knowledge. And, it defines those invariants as a set of mathematical equations, which can predict expected values of a sensor based on real observed data from another sensor. As comparing those predicted values and observed values, SIAT can distinguish abnormal behavior in the system. This comparison can be executed online to realize real-time anomaly detection and catch an early phenomenon of abnormal system behaviors. This detection cannot be easily conducted with typical rule-based detection method such as a threshold-based detection. With SIAT, the pattern of the comparison results in an invariant set will be able to help root cause localization after an anomaly detection.
NEC has been contributing to our society as providing innovative solutions for problems and difficulties in our life. We aim to realize a safe and secure society.
Steps for automatic anomaly detection
