Early-Stage Gas Identification Using Convolutional Long Short-Term Neural Network with Sensor Array Time Series Data

Gas identification/classification through pattern recognition techniques based on gas sensor arrays often requires the equilibrium responses or the full traces of time-series data of the sensor array. Leveraging upon the diverse gas sensing kinetics behaviors measured via the sensor array, a computational intelligence- based meta-model is proposed to automatically conduct the feature extraction and subsequent gas identification using time-series data during the transitional phase before reaching equilibrium. The time-series data contains implicit temporal dependency/correlation that is worth being characterized to enhance the gas identification performance and reliability. In this context, a tailored approach so-called convolutional long short-term memory (CLSTM) neural network is developed to perform the identification task incorporating temporal characteristics within time-series data. This novel approach shows the enhanced accuracy and robustness as compared to the baseline models, i.e., multilayer perceptron (MLP) and support vector machine (SVM) through the comprehensive statistical examination. Specifically, the classification accuracy of CLSTM reaches as high as 96%, regardless of the operating condition specified. More importantly, the excellent gas identification performance of CLSTM at early stages of gas exposure indicates its practical significance in future real-time applications. The promise of the proposed method has been clearly illustrated through both the internal and external validations in the systematic case investigation.

Experimental Analysis of Gas Holdup Measured by Gas Array Tool in Gas–Water Two Phase of Horizontal Well

In the gas-water two phase of horizontal well, gas holdup is usually obtained by inverse calculation of the water holdup measured by the array capacitance probes. Gas Array Tool (GAT) has been developed to directly measure gas holdup. This instrument has been introduced into China and its simulation experiment in gas-water two phase flow in horizontal wells has been carried out for the first time to study the applicability of gas holdup measurement. Firstly, the response principle and measurement method of GAT are analyzed; secondly, the experimental data of GAT under different flowrates, water cut, and different cable speed are plotted and analyzed; finally, the gas holdup data measured by GAT and Capacitance Array Tool (CAT) are compared by using an interpolation algorithm. It is found that the response of the optical fiber probe is consistent and stable. It also proves the accuracy of gas identification and the applicability of gas holdup measurement under test conditions by GAT, which lays a foundation for further gas holdup measurement, interpretation, and field test in the future.

Shallow Gas Identification using LWD Monopole Sonic: A-non Radioactive for Safe and Effective Logs Acquisition in Tunu Field, Mahakam Delta

The Tunu field is a swamp giant gas field located in the Mahakam Delta, East Kalimantan. Stratigraphically, this field has an anticline structure with three main intervals; Tunu Shallow Zone (TSZ), Fresh Water Zone (FWZ), and Tunu Main Zone (TMZ). Shallow gas reservoirs of TSZ have been produced since 2008, following the production of TMZ in the 1990s. Drilling targets in the shallow gas reservoir decreased significantly due to limited reservoir targets, high inclination wells and a low oil price environment. The utilization of radioactive source logging (density and neutron) on Logging While Drilling (LWD) tools is not recommended to be performed in open hole mode for operational and safety issues (e.g: tool stuck). Thus, LWD Monopole sonic is chosen as a replacement of LWD Neutron-Density logs and helps to differentiate between shallow gas potential and coal lithology which is the main challenge in TSZ at interval depth above 1200 mSS. The methodology utilized sonic semblance (STRA) and compressional slowness (DTc) data at real-time and memory data logs, so early decision can be made in drilling mode. In a gas-bearing reservoir, both semblance and slowness are missing, while in coal it produced strong semblance. In order to differentiate carbonate lithology, additional data, such as cutting, calcimetry, drilling Rate of Penetration and Gas While Drilling are utilized. During 2018-2020, 5 wells have been drilled using LWD Monopole sonic together with LWD GR-Resistivity-Neutron-Density (Triple Combo) to calibrate the fluid interpretation and 3 trial wells with only GR-Resistivity-Monopole Sonic. As a result, LWD Monopole sonic is able to differentiate between Gas and Coal based on semblance and slowness with a success ratio up to 80%. This LWD Monopole Sonic provides a non-radioactive solution for safe and effective logs acquisition for shallow gas identification that could be applied in oil and gas fields outside Mahakam.

Smart Multi-Sensor Platform for Analytics and Social Decision Support in Agriculture

Smart agriculture based on new types of sensors, data analytics and automation, is an important enabler for optimizing yields and maximizing efficiency to feed the world’s growing population while limiting environmental pollution. The aim of this paper is to describe a multi-sensor Internet of Things (IoT) system for agriculture consisting of a soil probe, an air probe and a smart data logger. The implementation details will focus of the integration element and the innovative Artificial Intelligence based gas identification sensor. Furthermore, the paper focuses on the analytics and decision support system implementation that provides farming recommendations and is enhanced with a feedback loop from farmers and a social trust index that will increase the reliability of the system.