True North Prognostics

Founded in 2025 in San Diego, True North Prognostics (TNP) is a technology think tank dedicated to using artificial intelligence (AI) and advanced analytics to proactively detect the incipience or onset of degradation mechanisms that can lead to costly asset failures and unplanned downtime for complex engineering assets across multiple industries. TNP creates digital solutions that increase plant and asset Reliability, Availability and Safety (RAS), and optimal energy efficiency, across defense, transportation, energy, data center, and advanced manufacturing sectors.

TNP’s goal is to harness the digital and AI revolution to build technologies that protect human life and the environment. The company is committed to innovation and collaboration, developing tools that anticipate problems before they occur and help industries operate more safely and efficiently.

At the heart of TNP’s work is the AI-MSET™(Artificial Intelligence Multivariate State Estimation Technique) — an advanced machine learning algorithm designed to detect early signs of mechanisms that can lead to system failures. AI-MSET™ learns how systems behave under normal conditions and continuously monitors sensor data to identify subtle changes that might signal the onset of developing anomalies in all types of engineering assets, or the onset of malicious intrusion activity for enhancement of prognostic cybersecurity in SCADA and other business-critical IT assets and networks.

Unlike traditional monitoring systems that rely on high/low thresholds applied to signals individually (i.e. univariate alerting), AI-MSET™ learns the patterns between and among any small or large collection of dynamically interacting sensors in real time. This allows AI-MSET™ to achieve very high prognostic sensitivity while keeping false alarms extremely low and at the same time the avoidance of missed alarms, even in noisy environments.
 

AI-MSET™ was approved by the US Nuclear Regulatory Commission (NRC) for monitoring all commercial nuclear plants in the US. It is also IEEE-published, peer-reviewed, and proven to outperform all other machine learning methods— such as LSTM and other neural networks and/or their derivative forms — in terms of accuracy, lead time for early warnings of the onset of anomalies, lowest false-alarm and missed-alarm probabilities, and  2-3  orders of magnitude lower compute cost (compared to neural networks monitoring the same numbers of signals). For multivariate time-series prognostics, competitive neural networks are limited to a few hundred signals, even on the most powerful GPUs [because neural nets do not parallelize on any multi-thread multi-core CPUs or GPUs, whereas MSET is deterministic math and achieves fine-grain parallelism on all modern chips]. For “big data” applications, TNP’s AI-MSET™ can monitor dense-sensor streaming analytics on inexpensive legacy CPUs that, for LSTM and other neural networks, require very costly GPU Clusters.

On smaller data sets comprised of dozens or hundreds of signals, AI-MSET™ can be deployed on low-cost and compact laptops, 1RU enterprise boxes, and Raspberry Pi’s.

TNP’s technologies are applied across many fields, including:

• Defense and Aerospace: Early detection of incipient faults in Defense assets, including land, air (fixed-wing and rotary-wing assets), sea (surface and underwater), and space; as well as commercial aircraft (including landing gear prognostics) and space systems.
• Transportation: Preventing train derailments, malicious intrusions in safety-critical switch & crossing (S&C) systems, and monitoring bridge, tie, and substrate integrity in real time.
• Situational Awareness: involving human-in-the-loop (HITL) supervisory control of complex engineering assets.
• High Frequency Waveforms: processing high-frequency waveforms on low-cost compute platforms and with substantially reduced I/O-bandwidth for sensors, including Vibrations (KHz), Acoustics (MHz), and IR Thermal Imagery (GHz).
• IT Systems: Electronic prognostics, optimized power management, software aging and rejuvenation, and prognostic cyber security.
• OT Systems: monitoring industrial environments to detect the early onset of malicious intrusion activity (and with ultra-low false-alarm rates).
• Manufacturing: Predictive and prescriptive maintenance that reduces downtime, scrap rates, Operation & Maintenance (O&M) costs, while enhancing quality for manufactured products, thereby reducing both warranty costs and sparing logistics for the manufacturer’s end customers.
• Energy: Monitoring all instrumentation and sensors for commercial nuclear power plants, improving wind turbine efficiency and RAS (reliability, availability, and serviceability), desalination efficiency, and lithium extraction.
• Healthcare and Safety: Detecting counterfeit pharmaceuticals and eliminating risks from thermal runaway in all classes of Lithium-Ion batteries.

In short, True North Prognostics is driving the future of predictive technology — merging AI, data science, and engineering to create smarter, safer, and more sustainable systems for industries that power our world.