| For a long time, mine gas disaster has been one of the most important factors threating the safety production of coal mine in China.The prevention and control of gas in heading face is the most important work in the prevention and control of mine gas disaster duringthe whole gas prevention and control process.At the same time, it may encounter all kinds of abnormal disaster in the tunneling working face excavation process, and the root causes of these abnormal disaster is the change of the coal body stress, gas pressure in coal and a variety of physical and mechanical properties in heading face of coal construction work;Andbecause of different types of tunneling working procedure, the damage degree of the stability of the coal body is different,which results in different types of tunneling processes with different gas emission regularity and characteristics, so it may cause the different types of abnormal gas disasters. Therefore, it is an important basis for the safety production and management decision of the coal mine to grasp the air flow and the law of gas emission in different types of operations. And it is also an important part of the mine ventilation design and the establishment of reasonable gas control measures.At the same time, the tunneling working surface gas sensors record the original monitoring data of gas concentration by time seriesevery day. Therefore, the type of real time working procedure type can be identified bythe characteristics of gas emission concentration in heading face, it also helps engineers to grasp more easily the intensity and judgment about the real time working condition and the strong influence of the front coal body on the working face of the driving working face. But in recent years, as a mathematical theory and a method of modern digital signal processing, wavelet transform has attracted more and more attention of experts and scholars in the fieldof science and technology.Wavelet analysis can be performed on the signal(function) by using the function of expansion and translation.In the time domain and frequency domain, it has the ability to hold the characteristic of signal, and it is a local transformation of time and frequency.And it can be focused to any detail of the signal and effectively extract the feature information from the signal. So wavelet analysis is a new signal processing technology.Especially wavelet packet analysis can provide a finer analysis approach for signal, it can decompose the low frequency and high frequency parts of the signal andcan be used for accurate feature extraction of nonstationary and abrupt signals.So it is more effective to reflect the time-frequency characteristics of the signal.Based on this, this paper first selects the Shanxi coal import and export group left in Hongyuan Coal Industry Co., Ltd.(hereinafter referred to as the Hongyuan coal mine)of 150201 fully mechanized working face return along the groove on the comprehensivemechanized mining faces and 150201 fully mechanized coal mining face transport Shun groove blasting driving face as experimental roadway developed field measurement scheme is used to field test the digging working face raw data, such as gas concentration,wind speed, air volume, footage and the excavation process of work starting time.Secondly based on the theory of coal seam gas occurrence and coal seam gas flow to establish mathematical model of different working procedure in the heading face gas emission. And thedifferent operation gas emission in working face was simulated, to verify the correctness of the mathematical model with comparing with the measured data, so as to determine the gas emission law of different roadway operation.At the same time simulating air flow field and gas ventilation heading face to get the general law of airflow field and gas distribution.Providing a reliable theoretical basis to prevent gas excavation face aggregation reasonable measures, so as to mine local ventilation to provide theoretical guidance and technical and ensure safety production.What is the concentration of gas emission time sequence data according to the tunneling working surface by field measurement scheme the original monitoring, using the wavelet analysis method to denoise the concentration time series data on the raw gas emission,and the use of wavelet packet decomposition to extract the heading face in the process of tunneling excavation operation different gas emission the time series of signal energy spectrum as the feature vector, finally using gas from the above wavelet denoising and wavelet packet transform to get the emission feature vector by fuzzy clustering model classifier to identify different types of driving operation driving method.The specific contents are as follows:(1) Determine the object of pattern recognition and specific identification methodFirstly, through the analysis and research on the theory of pattern recognition, and combining with the influence factors which determined by the type of roadway dynamic gas emission and the actual excavation work process determine the types of objects.This paper determined by time series of tunneling face gas emission concentration as the original monitoring signals and to all kinds of tunneling process as object.(2) The original gas concentration data acquisition by time seriesSetting the original data scheme.First select the measurement object(select outstanding dangerous blasting driving face and fully mechanized excavation face of a second), after set measuring point data acquisition such as gas concentration, wind speed, air volume, and the excavation process of excavation process start time and transformthe original data into the computer can process data.Providing comparative data basis about the tunneling face gas emission law and dynamic follow-up study on wavelet denoising and wavelet packet feature extraction.(3)Thelaw research about different heading face gas emission in work processApplying the flow of coal seam gas occurrence and coal seam gas theory toestablish the mathematical model of the tunneling working surface of different working gas emission and usingthe FLUENT software on the heading face in different operation process numerical simulation of gas emission law tocalculate the amount of gas emitted from the heading face. And by comparing with measured data to verify the correctness of the mathematical model, so as to determine the gas emission law of different roadway operation.At the same time, by simulating the air flow field and gas ventilation heading face, getting the general law of airflow field and gas distribution.It provided a reliable theoretical basis for preventing gas accumulation in heading face.(4) The pretreatmentabout original data and the gas emission feature extractionThe concentration data of gas emission time sequence driving face above by field measurement scheme based on the original monitoring.Firstly, using wavelet analysis denoising method of the original gas concentration time series data denoising preprocessing.At the same time, as the wavelet packet multi-resolution analysis in different scales with different time and frequency resolution, can separate the different frequency components.It also solves the different decomposition scales of energy. So we use wavelet packet transform on the gas concentration time series signal after de-noising decomposition extraction in the tunneling process of different excavation operation process of gas emission time series signal of each frequency band energy spectrum as the feature vector, and finally extract the various frequency bands of the energy spectrum feature vector as the input of fuzzy clustering pattern recognition.(5) Construct pattern classifier for pattern recognition and application verificationBy using above wavelet denoising and wavelet packet decomposition to extract the different driving operation characteristics of gas emission standard samples,and the feature vector and the standard sample concentration to establish sample characteristics of normalized index matrix, the application of absolute value method to calculate the correlation degree are reciprocal, fuzzy similarity matrix.Then the fuzzy equivalence relation matrix transitiveclosure method, according to the standard sample set to classify the intercept.Through the method of fuzzy clustering, the sample set is divided into several categories,to establish a base of fuzzy pattern of different types of driving operation.Finally, new samples to be identified by pattern recognition, to verify the validity of the pattern recognition and the accuracy of the method.Through the above wavelet analysis combined with fuzzy clustering pattern recognition pattern recognition of different types of driving operation,and on the heading face in different types of operation when the air flow and the gas flow law of coal seam was studied by numerical simulation.So that the coal mine can occur abnormal disaster ahead of the working face excavation process to take appropriate security measures to prevent disaster. So the disaster will be killed in the bud.This research has very important practical significance and application value to significantly improve the safety level of coal mine tunneling and improving the grim. |
PDF Full Text Request