Time-series gas prediction model using LS-SVR within a Bayesian framework

The traditional least squares support vector regression（LS-SVR）model,using cross validation to determine the regularization parameter and kernel parameter,is time-consuming.We propose a Bayesian evidence framework to infer the LS-SVR model parameters.Three levels Bayesian inferences are used to determine the model parameters,regularization hyper-parameters and tune the nuclear parameters by model comparison.On this basis,we established Bayesian LS-SVR time-series gas forecasting models and provide steps for the algorithm.The gas outburst data of a Hebi 10th mine working face is used to validate the model.The optimal embedding dimension and delay time of the time series were obtained by the smallest differential entropy method.Finally,within a MATLAB7.1 environment,we used actual coal gas data to compare the traditional LS-SVR and the Bayesian LS-SVR with LS-SVMlab1.5 Toolbox simulation.The results show that the Bayesian framework of an LS-SVR significantly improves the speed and accuracy of the forecast.

Cooperative prediction method of gas emission from mining face based on feature selection and machine learning

Collaborative prediction model of gas emission quantity was built by feature selection and supervised machine learning algorithm to improve the scientifc and accurate prediction of gas emission quantity in the mining face.The collaborative prediction model was screened by precision evaluation index.Samples were pretreated by data standardization,and 20 characteristic parameter combinations for gas emission quantity prediction were determined through 4 kinds of feature selection methods.A total of 160 collaborative prediction models of gas emission quantity were constructed by using 8 kinds of classical supervised machine learning algorithm and 20 characteristic parameter combinations.Determination coefcient,normalized mean square error,mean absolute percentage error range,Hill coefcient,mean absolute error,and the mean relative error indicators were used to verify and evaluate the performance of the collaborative forecasting model.As such,the high prediction accuracy of three kinds of machine learning algorithms and seven kinds of characteristic parameter combinations were screened out,and seven optimized collaborative forecasting models were fnally determined.Results show that the judgement coefcients,normalized mean square error,mean absolute percentage error,and Hill inequality coefcient of the 7 optimized collaborative prediction models are 0.969–0.999,0.001–0.050,0.004–0.057,and 0.002–0.037,respectively.The determination coefcient of the fnal prediction sequence,the normalized mean square error,the mean absolute percentage error,the Hill inequality coefcient,the absolute error,and the mean relative error are 0.998%,0.003%,0.022%,0.010%,0.080%,and 2.200%,respectively.The multi-parameter,multi-algorithm,multi-combination,and multijudgement index prediction model has high accuracy and certain universality that can provide a new idea for the accurate prediction of gas emission quantity.

Prediction method for risks of coal and gas outbursts based on spatial chaos theory using gas desorption index of drill cuttings

Based on the evolution of geological dynamics and spatial chaos theory, we proposed the advanced prediction an advanced prediction method of a gas desorption index of drill cuttings to predict coal and gas outbursts. We investigated and verified the prediction method by a spatial series data of a gas desorption index of drill cuttings obtained from the 113112 coal roadway at the Shitai Mine. Our experimental results show that the spatial distribution of the gas desorption index of drill cuttings has some chaotic charac- teristics, which implies that the risk of coal and gas outbursts can be predicted by spatial chaos theory. We also found that a proper amount of sample data needs to be chosen in order to ensure the accuracy and practical maneuverability of prediction. The relative prediction error is small when the prediction pace is chosen carefully. In our experiments, it turned out that the optimum number of sample points is 80 and the optimum prediction pace 30. The corresponding advanced prediction pace basically meets the requirements of engineering applications.

Application of computational intelligence platform in coal and gas outburst prediction

The present situation of lacking fast and effective coal and gas outburst prediction techniques will lead to long out- burst prevention cycles and poor accurate prediction effects and slows down coal roadway drive speed seriously. Also, due to historical and economic reasons, some coal mines in China are equipped with poor safety equipment, and the staff professional capability is low. What＇s worse, artificial and mine geological conditions have great influences on the traditional technologies of coal and gas outburst prediction. Therefore, seeking a new fast and efficient coal and gas outburst prediction method is nec- essary. By using system engineering theory, combined with the current mine production conditions and based on the coal and gas outburst composite hypothesis, a coal and gas outburst spatiotemporal forecasting system was established. This system can guide forecasting work schedule, optimize prediction technologies, carry out step-by-step prediction and eliminate hazard hier- archically. From the point of view of application, the proposed system improves the prediction efficiency and accuracy. On this basis, computational intelligence methods to construct disaster information analysis platform were used. Feed-back results pro- vide decision support to mine safety supervisors.

Control of facies and fluid potential on hydrocarbon accumulation and prediction of favorable Silurian targets in the Tazhong Uplift,Tarim Basin,China

Exploration practices show that the Silurian hydrocarbon accumulation in the Tazhong Uplift is extremely complicated.Our research indicates that the oil and gas accumulation is controlled by favorable facies and low fluid potential.At the macro level,hydrocarbon distribution in this uplift is controlled by structural zones and sedimentary systems.At the micro level,oil occurrences are dominated by lithofacies and petrophysical facies.The control of facies is embodied in high porosity and permeability controlling hydrocarbon accumulation.Besides,the macro oil and gas distribution in the uplift is also influenced by the relatively low fluid potential at local highs,where most successful wells are located.These wells are also closely related to the adjacent fractures.Therefore,the Silurian hydrocarbon accumulation mechanism in the Tazhong Uplift can be described as follows.Induced by structures,the deep and overpressured fluids migrated through faults into the sand bodies with relatively low potential and high porosity and permeability.The released overpressure expelled the oil and gas into the normal-pressured zones,and the hydrocarbon was preserved by the overlying caprock of poorly compacted Carboniferous and Permian mudstones.Such a mechanism reflects favorable facies and low potential controlling hydrocarbon accumulation.Based on the statistical analysis of the reservoirs and commercial wells in the uplift,a relationship between oil-bearing property in traps and the facies-potential index was established,and a prediction of two favorable targets was made.

Forecasting and optimal probabilistic scheduling of surplus gas systems in iron and steel industry

To make full use of the gas resource, stabilize the pipe network pressure, and obtain higher economic benefits in the iron and steel industry, the surplus gas prediction and scheduling models were proposed. Before applying the forecasting techniques, a support vector classifier was first used to classify the data, and then the filtering was used to create separate trend and volatility sequences. After forecasting, the Markov chain transition probability matrix was introduced to adjust the residual. Simulation results using surplus gas data from an iron and steel enterprise demonstrate that the constructed SVC-HP-ENN-LSSVM-MC prediction model prediction is accurate, and that the classification accuracy is high under different conditions. Based on this, the scheduling model was constructed for surplus gas operating, and it has been used to investigate the comprehensive measures for managing the operational probabilistic risk and optimize the economic benefit at various working conditions and implementations. It has extended the concepts of traditional surplus gas dispatching systems, and provides a method for enterprises to determine optimal schedules.

Forecast of Geological Gas Hazards for “Three-Soft” Coal Seams in Gliding Structural Areas

Gas outbursts from "three-soft" coal seams (soft roof,soft floor and soft coal) constitute a very serious prob-lem in the Ludian gliding structure area in western Henan. By means of theories and methods of gas geology,structural geology,coal petrology and rock tests,we have discussed the effect of control of several physical properties of soft roof on gas preservation and proposed a new method of forecasting gas geological hazards under open structural conditions. The result shows that the areas with type Ⅲ or Ⅳ soft roofs are the most dangerous areas where gas outburst most likely can take place. Therefore,countermeasures should be taken in these areas to prevent gas outbursts.

Predicting gas-bearing distribution using DNN based on multi-component seismic data: Quality evaluation using structural and fracture factors

The tight-fractured gas reservoir of the Upper Triassic Xujiahe Formation in the Western Sichuan Depression has low porosity and permeability. This study presents a DNN-based method for identifying gas-bearing strata in tight sandstone. First, multi-component composite seismic attributes are obtained.The strong nonlinear relationships between multi-component composite attributes and gas-bearing reservoirs can be constrained through a DNN. Therefore, we identify and predict the gas-bearing strata using a DNN. Then, sample data are fed into the DNN for training and testing. After optimized network parameters are determined by the performance curves and empirical formulas, the best deep learning gas-bearing prediction model is determined. The composite seismic attributes can then be fed into the model to extrapolate the hydrocarbon-bearing characteristics from known drilling areas to the entire region for predicting the gas reservoir distribution. Finally, we assess the proposed method in terms of the structure and fracture characteristics and predict favorable exploration areas for identifying gas reservoirs.

Prediction of blast furnace gas generation based on data quality improvement strategy

The real-time energy flow data obtained in industrial production processes are usually of low quality.It is difficult to accurately predict the short-term energy flow profile by using these field data,which diminishes the effect of industrial big data and artificial intelligence in industrial energy system.The real-time data of blast furnace gas(BFG)generation collected in iron and steel sites are also of low quality.In order to tackle this problem,a three-stage data quality improvement strategy was proposed to predict the BFG generation.In the first stage,correlation principle was used to test the sample set.In the second stage,the original sample set was rectified and updated.In the third stage,Kalman filter was employed to eliminate the noise of the updated sample set.The method was verified by autoregressive integrated moving average model,back propagation neural network model and long short-term memory model.The results show that the prediction model based on the proposed three-stage data quality improvement method performs well.Long short-term memory model has the best prediction performance,with a mean absolute error of 17.85 m3/min,a mean absolute percentage error of 0.21%,and an R squared of 95.17%.