Advanced Sensors and Algorithms for GCS Leak Detection

Ensuring safety and public acceptance of Geologic Carbon Sequestration GCS), a preferred means of mitigating CO2 emissions from fossil-fueled plants, requires novel cost-effective tools and methods for monitoring, verifying, and accounting MVA) to detect any CO2 leakage from sequestration reservoirs. Identifying leaks is challenging because they are difficult to distinguish from the varying natural and man-made ambient CO2. Requisite tools include reliable long-term sensitive, autonomous and cost- effective measurements and data processing algorithms to detect, locate, and quantify sources of any escaping CO2. Physical Sciences Inc. PSI) is developing cost effective laser-based gas monitoring sensors that continuously and autonomously monitor near-surface concentrations of CO2. Long open-path sensors operated continuously for months at test sites including the Midwest Geological Sequestration Consortiums MGSC) Illinois Basin Decatur Project IBDP). Field data acquired with open-path sensors reveal that CO2 leaks from pipeline infrastructure create distinct anomalies in the concentration temporal statistics. Leak plumes cause rapid fluctuations that are readily distinguished from relatively slow natural background variations e.g. diurnal fluctuations). We are currently developing laser-based sensors for deep-downhole and shallow sub-surface CO2 monitoring. The proposed project, through Phase II, will build and install at a GCS validation test site a network of shallow subsurface sensors that monitor short and long term statistical fluctuations of CO2 concentrations, thereby detecting any excess emissions indicating leakage from within the sequestration site. The work will include developing and refining advanced algorithms utilizing unsupervised classifiers that learn over time to recognize normal long-term statistical variations and discern abnormalities. These algorithms will automate processing of long-term data to recognize signatures of slow seepage from CO2. In Phase II, a suite of sensors will be installed at one or more validation test sites enabling quantitative measurements of flux and other statistical leakage indicators. Phase I will demonstrate feasibility by: 1) developing advanced statistics-based algorithms that automatically recognize statistical features of recorded sensor data indicative of abnormal leak) contributions to the observed CO2 levels; 2) applying the algorithms to post-process data previously acquired with the PSI open-path sensors; 3) complete the assembly of a prototype shallow subsurface sensor; and 4) test the updated algorithms while continuing new data acquisition with the sensors. These MVA sensing tools and algorithmic data-processing engines will immediately support ongoing GCS research projects. They will ultimately provide commercial benefits to GCS site operators by reducing the cost of leak surveying and verification of site integrity. There will also be commercial benefits, in the form of sensor sales, to the sensor manufacturer. The future market potential for such sensors exceeds $100M. Further, machine learning algorithms offer the potential for extension to other sensor types and applications under exploration, yielding enhanced sensitivities and commercially competitive advantage.