About Us

---

MSET (Multivariate State Estimation Technique) was originally developed at Argonne National Laboratory in the 1990s and subsequently licensed across hundreds of use cases for numerous government and commercial applications, including safety-critical, mission-critical, and high–return-on-investment business assets. The methodology has since been re-envisioned and advanced at TNP.

The original MSET framework constituted a statistically grounded pattern-recognition approach capable of rapid detection of incipient anomalies in multivariate sensor data from diverse classes of engineered assets. The technique was mathematically proven to achieve optimal performance in prognostic anomaly detection by simultaneously minimizing both false-alarm and missed-alarm probabilities.

TNP has extended the original MSET core algorithms through the development of autonomous AI embodiments, enhancing usability and applicability for large-scale time-series anomaly detection. These advancements enable scalability to arbitrarily large sensor populations, accommodate sampling rates spanning from hertz (Hz) to kilohertz (KHz), megahertz (MHz), and gigahertz (GHz), and achieve computational efficiencies that are two to three orders of magnitude greater than those of long short-term memory (LSTM) and other neural network–based approaches. TNP’s AI-MSET™ represents a significant evolution in prognostic intelligence, supporting asset health monitoring and predictive analytics across multiple industrial domains.