CN104182802A - Automatic and visual prediction method of coal seam methane content - Google Patents

Abstract

The invention relates to an automatic and visual prediction method of a coal seam gas-bearing capacity, a whole gas-bearing capacity prediction process is divided into six steps of obtaining original data of gas-bearing capacity; setting relative parameters required for modeling; selecting a grey model, a quantification theory or a neural network algorithm to establish a gas-bearing capacity multivariable prediction model automatically; selecting a gas-bearing capacity prediction method (click or batch prediction) to carry out interactive prediction on an unknown region; and displaying establishment and prediction results of the gas-bearing capacity model.

Description

A kind of coal seam gas-bearing capacity robotization and visual Forecasting Methodology

Technical field

The present invention relates to a kind of coal seam gas-bearing capacity robotization and visual Forecasting Methodology.

Background technology

Coal seam gas-bearing capacity is the amount of gas volume (being scaled standard state) contained in unit mass coal.It is a major parameter of research coal seam gas bearing situation, it is the important evidence of Mine Gas Emission Prediction, coal and gas outburst prevention treatment and gas drainage, also be the requisite data of Design of Mine Ventilation, be related to the reasonable solution of the series of problems such as development system, ventilating system and mode, ventilation equipment, coal winning method, main roadway layout.

Research and development is put into practice and is shown, the tax of gas bearing capacity is deposited and distributed within the specific limits, has certain distribution regularity.At present, in production real process, set up the quantitative relationship of gas bearing capacity and coal seam buried depth, utilize direct gradient method and indirect gradient method to become the main method of prediction deep gas bearing capacity.Although this method is simple, practical, because the variation of the conditions such as tectonic structure and cap rock lithology often causes the significant difference of gas bearing capacity gradient, only consider that single factor Gas content prediction method of buried depth has serious application limitation.

Given this, some scholars utilize the method for statistics, have set up multifactor gas bearing capacity linear prediction model.Soup friendship etc. is utilized coal mine excavation district coal seam gas-bearing capacity measured value, on coalbed gas geology qualitative analysis basis, sets up the applicable Gas Content Prediction in Coal Seam formula in unworked country, field with "nine squares"; Wang Sheng applies the association analysis method of gray system theory entirely, on affecting the geologic agent of coal seam gas-bearing capacity, analyzes, and has found out major control factor, has constructed gas bearing capacity regressive prediction model; Zhong Lingwen etc. have set up the Gas Content Prediction in Coal Seam method under original place Gas Reservoir Connditions of Coal (comprising temperature, pressure, moisture, ash content, coal metamorphism) combined influence that approaches; The sub-woods of fine strain of millet carries out regretional analysis by apparent resistivity in well-log information and gamma gamma with the coal sample gas bearing capacity that lab analysis goes out, and sets up regression equation, and then dopes other coal seam gas-bearing capacitys and contrast etc. with lab analysis result.

In many situations, relation between coal seam gas-bearing capacity and its influence factor is complicated, utilizes the linear equation inner link between them beyond expression of words, and this just need to adopt nonlinear Forecasting Methodology.Ye Qing, Lin Baiquan application gray system theory, set up grey systems GM (1, the 1) model of Forecast Coal Seam Gas Content, and with Remanent Model, forecast model revised, and then on the basis of measuring coal seam gas-bearing capacity, carried out practical application; Zhang Keshu etc. utilize gray system theory to carry out association analysis to affecting the factor of coal seam gas-bearing capacity, major influence factors and secondary cause have been found out, and (1, N) grey forecasting model has carried out the system prediction under multifactor impact to coal seam gas-bearing capacity to utilize GM; Cui Gang, Lian Chengbo etc. set up BP neural network and carry out Gas content prediction, and prove that the method is better than the method for multiple linear regression; Wu's wealth virtue, once brave general's neural network organically combined with genetic algorithm, take neural network theory as basic, utilized the connection weights in genetic algorithm optimization hidden layer neuron number and network, had set up Gas content prediction model, etc.

Although the precision of prediction of multivariable nonlinearity Gas content prediction model is higher, because model is complicated, colliery technician is difficult to grasp, and the application in colliery engineering practice is considerably less.Therefore, need to set up a kind of gas bearing capacity automatic prediction method, several multivariate Gas content prediction numerical models are provided, realize Gas content prediction process and predict the outcome visual, finally for colliery technician provides a kind of convenient, fast, directly perceived, reliable Gas content prediction new tool.

At present, in production real process, the method that colliery technician carries out deep or unworked country Gas content prediction is mainly single factor gradient method, and specific implementation process is as follows:

1) add up mine gas bearing capacity measured data in the past;

2) analyze the relation that affects of gas bearing capacity and coal seam buried depth;

3) set up the linear regression model (LRM) of gas bearing capacity and coal seam buried depth;

4) utilize the regression model of setting up to carry out Gas Content Prediction in Coal Seam.

Also there are the multiple linear regression of employing or nonlinear theory to set up Gas Content Prediction in Coal Seam model, but apply less, not general.

Coal-mine gas content is a parameter that affected by many geologic agents, present stage colliery generally adopt single factor gradient method to predict gas bearing capacity, the reliability predicting the outcome is not high, directly affects Safety of Coal Mine Production.

Some scholars adopt suitable modeling method to set up multivariate Gas content prediction model, have improved forecasting reliability.Researcher adds up gas bearing capacity data, and the various factors of analyzing influence gas bearing capacity filters out major control factor, sets up corresponding forecast model, and territory, unworked country is predicted, finally the list that predicts the outcome is shown.Due to Gas content prediction, to relate to data many, contain much information, so this forecasting process has the deficiency of four aspects with predicting the outcome: 1) analyses and prediction process expends a large amount of man power and materials, and due to the difference in understanding and the difference of subjective consciousness, predict the outcome and can vary with each individual; 2) the coalbed gas geology data of reflection predicted condition constantly change with coal production, but traditional predicting the outcome is static, can not upgrade in time along with the accumulation of coalbed gas geology data, so also just can not provide in time up-to-date, predict achievement the most accurately; 3) Forecasting Methodology is not easy to combine different affecting factors to analyze its correlativity; 4) data are separated with result (map), are not easy to the consistency checking of data and history analysis and the persistence of data; 5) forecasting process and result do not realize robotization, visual.

Summary of the invention

The invention provides a kind of Gas content prediction new method, graphical information and data message organic unity are got up, set up gas bearing capacity multivariable prediction model, realize forecasting process and the robotization predicting the outcome, visual, not only improved the reliability predicting the outcome, but also improved the efficiency of Gas content prediction and the science of decision-making, can be grasped rapidly by general colliery engineering technician.

The present invention relates to a kind of coal seam gas-bearing capacity robotization and visual Forecasting Methodology, described method comprises the steps:

Whole Gas content prediction process is divided into six steps: obtain gas bearing capacity raw data; Model is set up to required correlation parameter to be arranged; Select grey modeling, theory of quantification or neural network algorithm automatically to set up gas bearing capacity multivariable prediction model; Select Gas content prediction method (clicking or batch forecast) to carry out interactive mode prediction to zone of ignorance; Show that gas bearing capacity model is set up and the result of prediction.

Accompanying drawing explanation

By describing in more detail exemplary embodiment of the present invention with reference to accompanying drawing, above and other aspect of the present invention and advantage will become and more be readily clear of, in the accompanying drawings:

Fig. 1 is for realizing gas bearing capacity robotization and the visual prediction framework schematic diagram based on multivariable model;

Fig. 2 is grey modeling numerical algorithm program flow diagram;

Fig. 3 is theory of quantification I numerical algorithm program flow diagram;

Fig. 4 is BP neural net model establishing numerical algorithm program flow diagram.

Embodiment

Hereinafter, now with reference to accompanying drawing, the present invention is described more fully, various embodiment shown in the drawings.Yet the present invention can implement in many different forms, and should not be interpreted as being confined to embodiment set forth herein.On the contrary, it will be thorough with completely providing these embodiment to make the disclosure, and scope of the present invention is conveyed to those skilled in the art fully.

Hereinafter, exemplary embodiment of the present invention is described with reference to the accompanying drawings in more detail.

Whole Gas content prediction process is divided into six steps: obtain gas bearing capacity raw data; Model is set up to required correlation parameter to be arranged; Select certain algorithm (grey modeling, theory of quantification or neural network) automatically to set up gas bearing capacity multivariable prediction model; Select Gas content prediction method (clicking or batch forecast) to carry out interactive mode prediction to zone of ignorance; Show that gas bearing capacity model is set up and the result of prediction.

As shown in Figure 1, gas bearing capacity multivariable prediction model is set up needed raw data and is directly referred in digitizing gas-geologic map, predict the outcome and automatic powder adding is added on platform, modeling and prediction also contact with gas-geologic map platform by dynamic link library (dll), this just guarantees that whole model is set up and process and the map of prediction gas combine, the nonlinear operation of realization prediction and visual.In addition, according to predicting the outcome, can adjustment model parameter (as reselect gas bearing capacity influence factor, increase or delete gas bearing capacity raw data that modeling is required etc.), re-start modeling and prediction.

The present invention adopts graph processing technique, computer programming, database technology and GIS method, multivariate modeling and forecasting process and digitizing coalbed gas geology map are combined, realize the performance prediction of gas bearing capacity and visual, improve the efficiency of prediction and the science of decision-making, final colliery technician provides a kind of convenient, fast, directly perceived, reliable Gas content prediction new tool.

1, the dynamic link library of automatic prediction exploitation

The dynamic link library of gas bearing capacity multivariate modeling is the core of the visual prediction of whole robotization, and it has realized the numerical algorithm that model is set up and predicted, and the interface function providing by it is realized and gas-geologic map compilation platform carries out alternately.Patent of the present invention provides three kinds of dynamic link library: gstpredict.dll, quantipredict.dll and bpnetpredict.dll, has realized respectively the numerical algorithm of grey modeling, theory of quantification I and BP neural network.

1) grey modeling dynamic link library

Ash modeling dynamic link library has been realized original data processing, has constructed matrix of coefficients, has been solved the whole modeling process such as coefficient vector, and is undertaken alternately by interface function and gas-geologic map compilation platform.

In developed grey modeling dynamic link library (gstpredict.dll) file, following 2 grey modeling interface functions are provided.

(1) GMlN function

(2) GMON function

The program circuit that this dynamic link library is realized grey modeling numerical algorithm as shown in Figure 2.

2) theory of quantification I dynamic link library

The whole modeling process such as theory of quantification I dynamic link library has been realized original data processing, constructs matrix of coefficients, solved coefficient vector, testing accuracy, and undertaken alternately by interface function and gas-geologic map compilation platform.In developed theory of quantification I dynamic link library (quantipredict.dll) file, provide quantifyl interface function.

Its numerical algorithm program circuit of realizing theory of quantification I modeling of this dynamic link as shown in Figure 3.

3) BP neural network dynamic chained library

BP neural network dynamic chained library has been realized the normalization of sample data, various weights has been carried out to the whole modeling process such as initialization, error calculating, weights adjustment, and is undertaken alternately by interface function and gas-geologic map compilation platform.In this dynamic link library (bpnetpredict.dll) file, following 2 interface functions are provided.

(1) BpTrain function

(2) BpPredict function

Its numerical algorithm flow chart of realizing BP network modelling as shown in Figure 4.

The foregoing is only embodiments of the invention, be not limited to the present invention.The present invention can have various suitable changes and variation.All any modifications of doing within the spirit and principles in the present invention, be equal to replacement, improvement etc., within protection scope of the present invention all should be included in.

Claims (1)

1. coal seam gas-bearing capacity robotization and a visual Forecasting Methodology, is characterized in that:

Whole Gas content prediction process is divided into six steps: obtain gas bearing capacity raw data; Model is set up to required correlation parameter to be arranged; Select grey modeling, theory of quantification or neural network algorithm automatically to set up gas bearing capacity multivariable prediction model; Select Gas content prediction method (clicking or batch forecast) to carry out interactive mode prediction to zone of ignorance; Show that gas bearing capacity model is set up and the result of prediction.