Showing 1–2 of 2 results for author: Sheen, S

Ultrasonic and Electromagnetic Sensors for Downhole Reservoir Characterization

Authors: K. Wang, H. T. Chien, S. Liao, L. P. Yuan, S. H. Sheen, S. Bakhtiari, A. C. Raptis

Abstract: The current work covers the evaluation of ultrasonic and electromagnetic (EM) techniques applied to temperature measurement and flow characterization for Enhanced Geothermal System (EGS). We have evaluated both ultrasonic techniques and microwave radiometry for temperature gradient and profile measurements. A waveguide-based ultrasonic probe was developed to measure the temperature gradient. A sta… ▽ More The current work covers the evaluation of ultrasonic and electromagnetic (EM) techniques applied to temperature measurement and flow characterization for Enhanced Geothermal System (EGS). We have evaluated both ultrasonic techniques and microwave radiometry for temperature gradient and profile measurements. A waveguide-based ultrasonic probe was developed to measure the temperature gradient. A statistic approach on estimating the average grain size via spectral analysis of the scattered ultrasonic signals is introduced. For directional temperature measurement, different microwave antenna designs are compared numerically and an array loop antenna design is selected for further development. Finally techniques to characterize the porosity and permeability of a hot dry rock resource are presented. △ Less

Submitted 30 June, 2020; originally announced July 2020.

Comments: 7 pages, 5 figures

Journal ref: PROCEEDINGS, Thirty-Sixth Workshop on Geothermal Reservoir Engineering Stanford University, Stanford, California, January 31 - February 2, 2011 SGP-TR-191

Median Binary-Connect Method and a Binary Convolutional Neural Nework for Word Recognition

Authors: Spencer Sheen, Jiancheng Lyu

Abstract: We propose and study a new projection formula for training binary weight convolutional neural networks. The projection formula measures the error in approximating a full precision (32 bit) vector by a 1-bit vector in the l_1 norm instead of the standard l_2 norm. The l_1 projector is in closed analytical form and involves a median computation instead of an arithmatic average in the l_2 projector.… ▽ More We propose and study a new projection formula for training binary weight convolutional neural networks. The projection formula measures the error in approximating a full precision (32 bit) vector by a 1-bit vector in the l_1 norm instead of the standard l_2 norm. The l_1 projector is in closed analytical form and involves a median computation instead of an arithmatic average in the l_2 projector. Experiments on 10 keywords classification show that the l_1 (median) BinaryConnect (BC) method outperforms the regular BC, regardless of cold or warm start. The binary network trained by median BC and a recent blending technique reaches test accuracy 92.4%, which is 1.1% lower than the full-precision network accuracy 93.5%. On Android phone app, the trained binary network doubles the speed of full-precision network in spoken keywords recognition. △ Less

Submitted 7 November, 2018; originally announced November 2018.