THE OVERALL PROJECT OBJECTIVE IS TO DEVELOP AND DEMONSTRATE AN INTEGRATED SOLUTION FOR FAST AND ACCURATE DETECTION OF H2 LOSSES FROM H2-BLENDING NATURAL GAS (NG) PIPELINES DOWN TO 1 PPB LEVEL WITH HIGH SELECTIVITY AGAINST INTERFERENT GASSES (E.G., COMPONENTS OF NG) AND HIGH ROBUSTNESS AGAINST AMBIENT PARAMETER CHANGES. TO ACHIEVE THE OVERALL OBJECTIVE, THREE SPECIFIC OBJECTIVES ARE PROPOSED: 1. DEVELOP AN ELECTROCHEMICAL HYDROGEN PUMP THAT CAN BE INTEGRATED WITH A H2 SENSOR TO SELECTIVELY CONCENTRATE THE H2 BY OVER 5,000-FOLD (I.E., 1 PPM TO >5 PPM) AND PURIFY THE H2 IN THE SAMPLE AIR BEFORE PERFORMING H2 SENSING. 2. DEVELOP A RESISTIVE TYPE H2 SENSOR BASED ON ENGINEERED NANOSTRUCTURES THAT HAS A LIMIT OF DETECTION (LOD) AT 1 PPM, ACCURACY OF ±0.1 PPM, < 30 SECONDS RESPONSE TIME, < 60 SECONDS RECOVERY TIME AT < 80 OC TEMPERATURE WITH CONSTANT RH LEVELS. 3. DEVELOP A SUITE OF MACHINE LEARNING ALGORITHMS THAT CAN A) PERFORM QUANTIFICATION OF THE H2 CONCENTRATION WITH LIMIT OF DETECTION AT 10 PPB AND ACCURACY OF ± 5% H2 CONCENTRATION AT A 95% CONFIDENCE INTERVAL USING TRANSIENT RESPONSES OF THE H2 SENSOR, AND B) CORRECT THE BASELINE DRIFT IN THE INTEGRATED PUMP-SENSOR RESPONSE CAUSED BY FLUCTUATION OF AMBIENT CONDITIONS: -30~90 °C & 10%-98% RH.