scholarly journals Forecasting of Coalbed Methane Daily Production Based on T-LSTM Neural Networks

Symmetry ◽  
2020 ◽  
Vol 12 (5) ◽  
pp. 861
Author(s):  
Xijie Xu ◽  
Xiaoping Rui ◽  
Yonglei Fan ◽  
Tian Yu ◽  
Yiwen Ju
Keyword(s):  
Time Series ◽  
Transfer Learning ◽  
Coalbed Methane ◽  
Short Term Memory ◽  
Early Stage ◽  
Production Data ◽  
Production Forecast ◽  
Daily Production ◽  
Cbm Production ◽  
Lstm Network

Accurately forecasting the daily production of coalbed methane (CBM) is important forformulating associated drainage parameters and evaluating the economic benefit of CBM mining. Daily production of CBM depends on many factors, making it difficult to predict using conventional mathematical models. Because traditional methods do not reflect the long-term time series characteristics of CBM production, this study first used a long short-term memory neural network (LSTM) and transfer learning (TL) method for time series forecasting of CBM daily production. Based on the LSTM model, we introduced the idea of transfer learning and proposed a Transfer-LSTM (T-LSTM) CBM production forecasting model. This approach first uses a large amount of data similar to the target to pretrain the weights of the LSTM network, then uses transfer learning to fine-tune LSTM network parameters a second time, so as to obtain the final T-LSTM model. Experiments were carried out using daily CBM production data for the Panhe Demonstration Zone at southern Qinshui basin in China. Based on the results, the idea of transfer learning can solve the problem of insufficient samples during LSTM training. Prediction results for wells that entered the stable period earlier were more accurate, whereas results for types with unstable production in the early stage require further exploration. Because CBM wells daily production data have symmetrical similarities, which can provide a reference for the prediction of other wells, so our proposed T-LSTM network can achieve good results for the production forecast and can provide guidance for forecasting production of CBM wells.

scholarly journals Testing RNN-LSTM Forecasting with Simulated Astronomical Lightcurves

2019 ◽  
Author(s):  
Nachiketa Chakraborty
Keyword(s):  
Time Series ◽  
Short Term Memory ◽  
Flicker Noise ◽  
Temporal Structure ◽  
Batch Size ◽  
Electromagnetic Spectrum ◽  
X Rays ◽  
Optimal Test ◽  
Noise Type ◽  
Lstm Network

With an explosion of data in the near future, from observatories spanning from radio to gamma-rays, we have entered the era of time domain astronomy. Historically, this field has been limited to modeling the temporal structure with time-series simulations limited to energy ranges blessed with excellent statistics as in X-rays. In addition to ever increasing volumes and variety of astronomical lightcurves, there's a plethora of different types of transients detected not only across the electromagnetic spectrum, but indeed across multiple messengers like counterparts for neutrino and gravitational wave sources. As a result, precise, fast forecasting and modeling the lightcurves or time-series will play a crucial role in both understanding the physical processes as well as coordinating multiwavelength and multimessenger campaigns. In this regard, deep learning algorithms such as recurrent neural networks (RNNs) should prove extremely powerful for forecasting as it has in several other domains. Here we test the performance of a very successful class of RNNs, the Long Short Term Memory (LSTM) algorithms with simulated lightcurves. We focus on univariate forecasting of types of lightcurves typically found in active galactic nuclei (AGN) observations. Specifically, we explore the sensitivity of training and test losses to key parameters of the LSTM network and data characteristics namely gaps and complexity measured in terms of number of Fourier components. We find that typically, the performances of LSTMs are better for pink or flicker noise type sources. The key parameters on which performance is dependent are batch size for LSTM and the gap percentage of the lightcurves. While a batch size of $10-30$ seems optimal, the most optimal test and train losses are under $10 \%$ of missing data for both periodic and random gaps in pink noise. The performance is far worse for red noise. This compromises detectability of transients. The performance gets monotonically worse for data complexity measured in terms of number of Fourier components which is especially relevant in the context of complicated quasi-periodic signals buried under noise. Thus, we show that time-series simulations are excellent guides for use of RNN-LSTMs in forecasting.

Deep Network based on Long Short-Term Memory for Time Series Prediction of Microclimate Data inside the Greenhouse

2020 ◽  
Vol 19 (02) ◽  
pp. 2050013
Author(s):  
Sawsan Morkos Gharghory
Keyword(s):  
Time Series ◽  
Relative Humidity ◽  
Time Series Data ◽  
Short Term Memory ◽  
Weather Data ◽  
Series Data ◽  
Short Term ◽  
Term Memory ◽  
Long Short Term Memory ◽  
Lstm Network

An enhanced architecture of recurrent neural network based on Long Short-Term Memory (LSTM) is suggested in this paper for predicting the microclimate inside the greenhouse through its time series data. The microclimate inside the greenhouse largely affected by the external weather variations and it has a great impact on the greenhouse crops and its production. Therefore, it is a massive importance to predict the microclimate inside greenhouse as a preceding stage for accurate design of a control system that could fulfill the requirements of suitable environment for the plants and crop managing. The LSTM network is trained and tested by the temperatures and relative humidity data measured inside the greenhouse utilizing the mathematical greenhouse model with the outside weather data over 27 days. To evaluate the prediction accuracy of the suggested LSTM network, different measurements, such as Root Mean Square Error (RMSE) and Mean Absolute Error (MAE), are calculated and compared to those of conventional networks in references. The simulation results of LSTM network for forecasting the temperature and relative humidity inside greenhouse outperform over those of the traditional methods. The prediction results of temperature and humidity inside greenhouse in terms of RMSE approximately are 0.16 and 0.62 and in terms of MAE are 0.11 and 0.4, respectively, for both of them.

scholarly journals An approach towards data change rule based classification of driving maneuver with LSTM network

2020 ◽  
Author(s):  
Supriya Sarker ◽  
Md Mokammel Haque
Keyword(s):  
Time Series ◽  
Sensor Fusion ◽  
Short Term Memory ◽  
Extraction Process ◽  
Test Accuracy ◽  
Fusion Data ◽  
Time Series Dataset ◽  
Lstm Network ◽  
Multi Class Classification

The proposed work develops a Long Short Term Memory (LSTM) model for multi class classification of driving maneuver from sensor fusion time series dataset. The work also analyzes the significance of sensor fusion data change rule and utilized the idea with deep learning time series multi class classification of driving maneuver. We also proposed some hypotheses which are proven by the experimental results. The proposed model provides Train Accuracy: 99.98, Test Accuracy: 97.2021, Precision: 0.974848, Recall: 0.960154 and F1 score: 0.967028. The Mean Per Class Error (MPCE) is 0.01386. The significant rules can accelerate the feature extraction process of driving data. Moreover, it helps in automatic labeling of unlabeled dataset. Our future approach is to develop a tool for generating categorical label for unlabeled dataset. Besides, we have plan to optimize the proposed classifier using grid search. <br>

scholarly journals A Multivariate Long Short-Term Memory Neural Network for Coalbed Methane Production Forecasting

Symmetry ◽  
2020 ◽  
Vol 12 (12) ◽  
pp. 2045
Author(s):  
Xijie Xu ◽  
Xiaoping Rui ◽  
Yonglei Fan ◽  
Tian Yu ◽  
Yiwen Ju
Keyword(s):  
Neural Network ◽  
Coalbed Methane ◽  
Short Term Memory ◽  
Historical Data ◽  
Short Term ◽  
Term Memory ◽  
Production Forecasting ◽  
Long Short Term Memory ◽  
Cbm Production ◽  
Coalbed Methane Production

Owing to the importance of coalbed methane (CBM) as a source of energy, it is necessary to predict its future production. However, the production process of CBM is the result of the interaction of many factors, making it difficult to perform accurate simulations through mathematical models. We must therefore rely on the historical data of CBM production to understand its inherent features and predict its future performance. The objective of this paper is to establish a deep learning prediction method for coalbed methane production without considering complex geological factors. In this paper, we propose a multivariate long short-term memory neural network (M-LSTM NN) model to predict CBM production. We tested the performance of this model using the production data of CBM wells in the Panhe Demonstration Area in the Qinshui Basin of China. The production of different CBM wells has similar characteristics in time. We can use the symmetric similarity of the data to transfer the model to the production forecasting of different CBM wells. Our results demonstrate that the M-LSTM NN model, utilizing the historical yield data of CBM as well as other auxiliary information such as casing pressures, water production levels, and bottom hole temperatures (including the highest and lowest temperatures), can predict CBM production successfully while obtaining a mean absolute percentage error (MAPE) of 0.91%. This is an improvement when compared with the traditional LSTM NN model, which has an MAPE of 1.14%. In addition to this, we conducted multi-step predictions at a daily and monthly scale and obtained similar results. It should be noted that with an increase in time lag, the prediction performance became less accurate. At the daily level, the MAPE value increased from 0.24% to 2.09% over 10 successive days. The predictions on the monthly scale also saw an increase in the MAPE value from 2.68% to 5.95% over three months. This tendency suggests that long-term forecasts are more difficult than short-term ones, and more historical data are required to produce more accurate results.

Production-Data Analysis of Single-Phase (Gas) Coalbed-Methane Wells

2007 ◽  
Vol 10 (03) ◽  
pp. 312-331 ◽  
Author(s):  
Christopher R. Clarkson ◽  
R. Marc Bustin ◽  
John P. Seidle
Keyword(s):  
Coalbed Methane ◽  
Single Phase ◽  
Material Balance ◽  
Production Data ◽  
Reservoir Properties ◽  
Type Curve ◽  
Well Production ◽  
Single Well ◽  
Permeability Anisotropy ◽  
Cbm Production

Summary Coalbed-methane (CBM) reservoirs commonly exhibit two-phase-flow (gas plus water) characteristics; however, commercial CBM production is possible from single-phase (gas) coal reservoirs, as demonstrated by the recent development of the Horseshoe Canyon coals of western Canada. Commercial single-phase CBM production also occurs in some areas of the low-productivity Fruitland Coal, south-southwest of the high-productivity Fruitland Coal Fairway in the San Juan basin, and in other CBM-producing basins of the continental United States. Production data of single-phase coal reservoirs may be analyzed with techniques commonly applied to conventional reservoirs. Complicating application, however, is the unique nature of CBM reservoirs; coal gas-storage and -transport mechanisms differ substantially from conventional reservoirs. Single-phase CBM reservoirs may also display complex reservoir behavior such as multilayer characteristics, dual-porosity effects, and permeability anisotropy. The current work illustrates how single-well production-data-analysis (PDA) techniques, such as type curve, flowing material balance (FMB), and pressure-transient (PT) analysis, may be altered to analyze single-phase CBM wells. Examples of how reservoir inputs to the PDA techniques and subsequent calculations are modified to account for CBM-reservoir behavior are given. This paper demonstrates, by simulated and field examples, that reasonable reservoir and stimulation estimates can be obtained from PDA of CBM reservoirs only if appropriate reservoir inputs (i.e., desorption compressibility, fracture porosity) are used in the analysis. As the field examples demonstrate, type-curve, FMB, and PT analysis methods for PDA are not used in isolation for reservoir-property estimation, but rather as a starting point for single-well and multiwell reservoir simulation, which is then used to history match and forecast CBM-well production (e.g., for reserves assignment). CBM reservoirs have the potential for permeability anisotropy because of their naturally fractured nature, which may complicate PDA. To study the effects of permeability anisotropy upon production, a 2D, single-phase, numerical CBM-reservoir simulator was constructed to simulate single-well production assuming various permeability-anisotropy ratios. Only large permeability ratios (&gt;16:1) appear to have a significant effect upon single-well production characteristics. Multilayer reservoir characteristics may also be observed with CBM reservoirs because of vertical heterogeneity, or in cases where the coals are commingled with conventional (sandstone) reservoirs. In these cases, the type-curve, FMB, and PT analysis techniques are difficult to apply with confidence. Methods and tools for analyzing multilayer CBM (plus sand) reservoirs are presented. Using simulated and field examples, it is demonstrated that unique reservoir properties may be assigned to individual layers from commingled (multilayer) production in the simple two-layer case. Introduction Commercial single-phase (gas) CBM production has been demonstrated recently in the Horseshoe Canyon coals of western Canada (Bastian et al. 2005) and previously in various basins in the US; there is currently a need for advanced PDA techniques to assist with evaluation of these reservoirs. Over the past several decades, significant advances have been made in PDA of conventional oil and gas reservoirs [select references include Fetkovich (1980), Fetkovich et al. (1987), Carter (1985), Fraim and Wattenbarger (1987), Blasingame et al. (1989, 1991), Palacio and Blasingame (1993), Fetkovich et al. (1996), Agarwal et al. (1999), and Mattar and Anderson (2003)]. These modern methods have greatly enhanced the engineers' ability to obtain quantitative information about reservoir properties and stimulation/damage from data that are gathered routinely during the producing life of a well, such as production data and, in some instances, flowing pressure information. The information that may be obtained from detailed PDA includes oil or gas in place (GIP), permeability-thickness product (kh), and skin (s), and this can be used to supplement information obtained from other sources such as PT analysis, material balance, and reservoir simulation.

scholarly journals Time series analysis and long short-term memory (LSTM) network prediction of BPV current density

2021 ◽  
Vol 14 (4) ◽  
pp. 2408-2418 ◽  
Author(s):  
Tonny I. Okedi ◽  
Adrian C. Fisher
Keyword(s):  
Time Series ◽  
Current Density ◽  
Short Term Memory ◽  
High Accuracy ◽  
Seasonal Component ◽  
Short Term ◽  
Term Memory ◽  
Network Prediction ◽  
Long Short Term Memory ◽  
Lstm Network

LSTM networks are shown to predict the seasonal component of biophotovoltaic current density and photoresponse to high accuracy.

scholarly journals Predicting COVID-19 cases using bidirectional LSTM on multivariate time series

2021 ◽  
Author(s):  
Ahmed Ben Said ◽  
Abdelkarim Erradi ◽  
Hussein Ahmed Aly ◽  
Abdelmonem Mohamed
Keyword(s):  
Time Series ◽  
Clustering Algorithm ◽  
Short Term Memory ◽  
Multivariate Time Series ◽  
Health Sector ◽  
Cumulative Number ◽  
Disease Propagation ◽  
Proposed Model ◽  
Bidirectional Lstm ◽  
Lstm Network

AbstractTo assist policymakers in making adequate decisions to stop the spread of the COVID-19 pandemic, accurate forecasting of the disease propagation is of paramount importance. This paper presents a deep learning approach to forecast the cumulative number of COVID-19 cases using bidirectional Long Short-Term Memory (Bi-LSTM) network applied to multivariate time series. Unlike other forecasting techniques, our proposed approach first groups the countries having similar demographic and socioeconomic aspects and health sector indicators using K-means clustering algorithm. The cumulative case data of the clustered countries enriched with data related to the lockdown measures are fed to the bidirectional LSTM to train the forecasting model. We validate the effectiveness of the proposed approach by studying the disease outbreak in Qatar and the proposed model prediction from December 1st until December 31st, 2020. The quantitative evaluation shows that the proposed technique outperforms state-of-art forecasting approaches.

scholarly journals Fault Detection and Diagnosis Using Combined Autoencoder and Long Short-Term Memory Network

Sensors ◽  
2019 ◽  
Vol 19 (21) ◽  
pp. 4612 ◽  
Author(s):  
Pangun Park ◽  
Piergiuseppe Di Marco ◽  
Hyejeon Shin ◽  
Junseong Bang
Keyword(s):  
Time Series ◽  
Fault Detection ◽  
Short Term Memory ◽  
Fault Detection And Diagnosis ◽  
Short Term ◽  
Term Memory ◽  
Long Short Term Memory ◽  
Detection And Diagnosis ◽  
Lstm Network ◽  
Types Of Faults

Fault detection and diagnosis is one of the most critical components of preventing accidents and ensuring the system safety of industrial processes. In this paper, we propose an integrated learning approach for jointly achieving fault detection and fault diagnosis of rare events in multivariate time series data. The proposed approach combines an autoencoder to detect a rare fault event and a long short-term memory (LSTM) network to classify different types of faults. The autoencoder is trained with offline normal data, which is then used as the anomaly detection. The predicted faulty data, captured by autoencoder, are put into the LSTM network to identify the types of faults. It basically combines the strong low-dimensional nonlinear representations of the autoencoder for the rare event detection and the strong time series learning ability of LSTM for the fault diagnosis. The proposed approach is compared with a deep convolutional neural network approach for fault detection and identification on the Tennessee Eastman process. Experimental results show that the combined approach accurately detects deviations from normal behaviour and identifies the types of faults within the useful time.

scholarly journals Forecasting adverse surgical events using self-supervised transfer learning for physiological signals

2021 ◽  
Vol 4 (1) ◽  
Author(s):  
Hugh Chen ◽  
Scott M. Lundberg ◽  
Gabriel Erion ◽  
Jerry H. Kim ◽  
Su-In Lee
Keyword(s):  
Time Series ◽  
Transfer Learning ◽  
Short Term Memory ◽  
Computational Cost ◽  
Physiological Signals ◽  
Short Term ◽  
Physiological Signal ◽  
Data Phase ◽  
Unseen Data ◽  
Long Short Term Memory

AbstractHundreds of millions of surgical procedures take place annually across the world, which generate a prevalent type of electronic health record (EHR) data comprising time series physiological signals. Here, we present a transferable embedding method (i.e., a method to transform time series signals into input features for predictive machine learning models) named PHASE (PHysiologicAl Signal Embeddings) that enables us to more accurately forecast adverse surgical outcomes based on physiological signals. We evaluate PHASE on minute-by-minute EHR data of more than 50,000 surgeries from two operating room (OR) datasets and patient stays in an intensive care unit (ICU) dataset. PHASE outperforms other state-of-the-art approaches, such as long-short term memory networks trained on raw data and gradient boosted trees trained on handcrafted features, in predicting six distinct outcomes: hypoxemia, hypocapnia, hypotension, hypertension, phenylephrine, and epinephrine. In a transfer learning setting where we train embedding models in one dataset then embed signals and predict adverse events in unseen data, PHASE achieves significantly higher prediction accuracy at lower computational cost compared to conventional approaches. Finally, given the importance of understanding models in clinical applications we demonstrate that PHASE is explainable and validate our predictive models using local feature attribution methods.

Sign in / Sign up

Export Citation Format

Share Document