SBIR Phase II: Early Detection of Anomalies in Large-Scale Gas Networks

Eigenpatterns Inc., San Jose CA

Investigators

Abstract

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase II project is to reduce the incidence of natural gas pipeline failures across the country within the next 3 to 5 years. Every year there are a few hundred 'significant' pipeline accidents (fatalities or significant property damage) causing massive damage to life and property, dispersing hazardous materials and disrupting gas delivery services. Such events can have severe consequences for a utility including bankruptcy, billions of dollars in liabilities, civil/criminal penalties and higher insurance premiums. This proposed scalable and economical capability to proactively identify potential issues prior to catastrophic failure will significantly reduce the likelihood of such failures without requiring additional infrastructure. The project will also help prevent unauthorized third-party activity near pipelines, a leading cause of accidents. The methods developed in this project can be directly applied to improve detection accuracy in other contexts such as power-grids, computer cluster management and financial fraud detection. This SBIR Phase II project proposes to detect anomalies in large-scale gas-utility networks through statistical inference from continuously observed time-series data on pressure, temperature, and network characteristics. Anomalies within gas-utility networks occur for various reasons, such as sulphur or ice buildup, regulator malfunction, corrosion/aging of hardware, and human error. Such failures are often preceded by detectable signatures in the time-series of gas-pressure data. Early detection of such signatures with significant advance warning (90 minutes or more) allows corrective action that will avoid loss of life, property damage and service disruption. This project proposes new methods for the rapid estimation of short and medium timescale models of gas pressure behavior from real-time pipeline data, along with methods for constructing prediction bands through Monte Carlo and stochastic optimization techniques. Such methods are non-generic and their success relies crucially on exploiting specific structural properties unique to network-level gas-pressure time series. The proposed stochastic optimization techniques will probabilistically classify identified anomalies into failure type, allowing the prioritizing of network level emergency operations. The project will also develop scalable automated high-dimensional classification models to detect construction activity from satellite and other image data, to initiate preventive action. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.