神经网络在瓦斯光谱传感系统中的应用研究
|
摘要
煤矿瓦斯监测是我国当前煤矿安全生产监督管理的重中之重。开发瓦斯灾害实时监测技术及装备,是预防瓦斯事故的重要防线和保障措施。用激光光谱法并结合神经网络对瓦斯进行预警是一种全新的方法,该研究适合我国煤矿瓦斯监测预警要求的在线、实时、快速的系统,使得该研究在世界上占有一席之地,建立出可行的预警模型,从而能极大地缓解我国煤炭生产百万吨死亡率远高于国际标准的局面。
本文首先论述了基于可调谐半导体激光吸收光谱技术的煤矿瓦斯预警模型及相关理论。通过对气体近红外选择性吸收的理论分析,给出了气体吸收测量的理论依据,并确定了甲烷气体的吸收谱线。采用DFB激光器在控制电路下扫描1.65μm附近的一条甲烷振转吸收谱线,通过甲烷2ν3带R(3)线的吸收对气体浓度进行检测,对激光的波长调制,研究其最佳参数,检测二次谐波以获得更高的探测信噪比,建立出光谱数据的反演方法。
其次介绍了神经网络和BP算法的相关理论,提出了利用BP神经网络进行瓦斯浓度分析和预测建模的方法。利用神经网络通过对历史浓度数据进行学习,找出瓦斯浓度变化的内在规律,并将其存储在网络具体的权值和偏置值中,用以预测未来的数据。采用三层前馈神经网络对瓦斯浓度建立预测模型,讨论了网络的各种参数设置的问题。用Matlab进行了仿真实验,实验表明BP神经网络用于瓦斯浓度的预警是可行的。
最后,从硬件结构的优化角度入手,对Sigmoid函数的硬件实现方式进行了比较研究,找出了较合理的设计方案。利用EDA技术,采用自顶向下的设计方法,通过FPGA硬件技术实现了BP神经网络算法。
The coal mine gas monitoring is the most important for our country current coal mine safety production supervision management. The development of real-time gas disaster monitoring technology and equipment for preventing gas accidents is an important line of defense and safeguard measures. Using laser Spectrographic combined with neural network for early warning of gas is a brand-new method, which suits the early warning requirements of online, real-time, the fast system of coal mine gas monitoring in our country. It makes this study some standing in the world and establishing the feasible early warning model, and then it can alleviate the aspect that our country coal production megaton mortality rates are much higher than the international standard.
First, this article discusses coal mine gas early warning model and correlation theories which are based on tunable diode laser absorption spectrum technology. From theoretical analysis of the near-infrared selective absorption of gas, it gives the gas absorption survey theory basis and determined the methane absorption lines. Controlled by circuit, it uses DFB lasers to scan a vibration-rotation absorption line at 1.65μm nearby of methane. Through examining of the absorption of methane 2υ3 band R (3) line, it surveys the methane gas density. With wavelength modulation, it studies the optimum parameter. Detecting the second-harmonic to obtain a higher SNR, it sets up inversion method of spectrum data.
Next, the article introduces some correlation theories of neural network and the BP algorithm, and proposes the methods of gas density analysis and forecast modeling using the BP neural network. Neural network can learn from the historical density data and discover the inherent changing laws, and store in the network specific weights and bias values to predict the future data. Using three-layered feed forward neural network, it sets up gas density prediction model and discusses various parameters setting problems of network. Using Matlab to do simulation experiments, it shows that BP neural network is feasible for early warning of gas density.
Finally, from the optimization perspective of hardware structure, with comparative studying hardware realizations of sigmoid function, this article discovers the reasonable design proposal. Using EDA techniques and top-down design methodology, it achieves the BP neural network algorithm in hardware by FPGA technology.
本文首先论述了基于可调谐半导体激光吸收光谱技术的煤矿瓦斯预警模型及相关理论。通过对气体近红外选择性吸收的理论分析,给出了气体吸收测量的理论依据,并确定了甲烷气体的吸收谱线。采用DFB激光器在控制电路下扫描1.65μm附近的一条甲烷振转吸收谱线,通过甲烷2ν3带R(3)线的吸收对气体浓度进行检测,对激光的波长调制,研究其最佳参数,检测二次谐波以获得更高的探测信噪比,建立出光谱数据的反演方法。
其次介绍了神经网络和BP算法的相关理论,提出了利用BP神经网络进行瓦斯浓度分析和预测建模的方法。利用神经网络通过对历史浓度数据进行学习,找出瓦斯浓度变化的内在规律,并将其存储在网络具体的权值和偏置值中,用以预测未来的数据。采用三层前馈神经网络对瓦斯浓度建立预测模型,讨论了网络的各种参数设置的问题。用Matlab进行了仿真实验,实验表明BP神经网络用于瓦斯浓度的预警是可行的。
最后,从硬件结构的优化角度入手,对Sigmoid函数的硬件实现方式进行了比较研究,找出了较合理的设计方案。利用EDA技术,采用自顶向下的设计方法,通过FPGA硬件技术实现了BP神经网络算法。
The coal mine gas monitoring is the most important for our country current coal mine safety production supervision management. The development of real-time gas disaster monitoring technology and equipment for preventing gas accidents is an important line of defense and safeguard measures. Using laser Spectrographic combined with neural network for early warning of gas is a brand-new method, which suits the early warning requirements of online, real-time, the fast system of coal mine gas monitoring in our country. It makes this study some standing in the world and establishing the feasible early warning model, and then it can alleviate the aspect that our country coal production megaton mortality rates are much higher than the international standard.
First, this article discusses coal mine gas early warning model and correlation theories which are based on tunable diode laser absorption spectrum technology. From theoretical analysis of the near-infrared selective absorption of gas, it gives the gas absorption survey theory basis and determined the methane absorption lines. Controlled by circuit, it uses DFB lasers to scan a vibration-rotation absorption line at 1.65μm nearby of methane. Through examining of the absorption of methane 2υ3 band R (3) line, it surveys the methane gas density. With wavelength modulation, it studies the optimum parameter. Detecting the second-harmonic to obtain a higher SNR, it sets up inversion method of spectrum data.
Next, the article introduces some correlation theories of neural network and the BP algorithm, and proposes the methods of gas density analysis and forecast modeling using the BP neural network. Neural network can learn from the historical density data and discover the inherent changing laws, and store in the network specific weights and bias values to predict the future data. Using three-layered feed forward neural network, it sets up gas density prediction model and discusses various parameters setting problems of network. Using Matlab to do simulation experiments, it shows that BP neural network is feasible for early warning of gas density.
Finally, from the optimization perspective of hardware structure, with comparative studying hardware realizations of sigmoid function, this article discovers the reasonable design proposal. Using EDA techniques and top-down design methodology, it achieves the BP neural network algorithm in hardware by FPGA technology.
引文
[1]周孟然,李振璧,张银.煤矿瓦斯光谱信号探测的数据处理[J].煤炭科学技术,2006,34(9):45~48
[2]Zhou Mengran, Li Zhenbi. Research on coal-mine gas monitoring system controlled by annealing simulating algorithm[J]. Proceedings of SPIE 3rd International Symposium on Advanced Optical Manufacturing and Testing Technologies, 2007, 6723:27-1~27-8
[3]Zhou Mengran, Li Zhenbi, Zhu Zongjiu. Research of Examination Technology of the Second-harmonic in Coal Mine Gas Detector[J]. Proceedings of 2007 8th International Conference on Electronic Measurement & Instruments, 2007, 4:587~591
[4]陈祥光,裴旭东.神经网络技术及其应用[M].北京:中国电力出版社.2003
[5]戴葵.神经网络实现技术[M].长沙:国防科技大学出版社.1998
[6]赵不贿.在系统可编程器件与开发技术[M].北京:机械工业出版社.2004
[7]赵岷.基于FPGA的控制器研究[D].西安:西北工业大学.2002
[8]W. Jin, G. Stewart, B. Culshaw. Absorption Measurement of Methane Gas Using a Broad Light Source and Interferometric Signal Processing[J]. Optics Letters, 1993, 18(16):337~339
[9]H. Inaba, T. Kobayasi, M. Hirama. Optical-Fiber Network System for Air-Pollution Monitoring over a Wide Area by Optical Absorption Method[J]. Electronics Letters, 1979, (23):749~751
[10]J. P. Dakin, C. A. Wade, D. Pinchbeck. A Novel Optical Fiber Methane Sensor [J]. Proc.SPIE, 1987, 734(5):187~190
[11]H. Tai. Long Distance Simultaneous Detection of Methane and Acetylene by Using Diode Lasers Coupled with Optical Fibers[J]. IEEE Photonics Technology Letters, 1992, 4(7): 804~807
[12]F. D Amato, P. Mazzinghi, F. Castagnoli. Methane analyzer based on TDL’s for measurements in the lower stratosphere: design and laboratory tests[J]. Applied Physics, 2002, B 75:195~202
[13]Koichi Ichimura et al. Fiber optic methane gas detection system[J]. Hitachi Cable Review, 2006, 25:11~15
[14]S. Hunsmann et al. Absolute, high resolution water transpiration rate measurements on single plant leaves via tunable diode laser absorption spectroscopy(TDLAS) at 1.37μm[J]. Applied Physics, 2008, B 92:393~401
[15]A. Fried et al. Tunable infrared laser instruments for airborne atmospheric studies[J]. Applied Physics, 2008, B 92:409~417
[16]J. P. Silveira, F. Grasdepot. CH4 Optical Sensor Using a 1.31ìm DFB laser dioxde[J]. Sensors and Actuators, 1995, 25(8):603~606
[17]T. Nakaya, K. Kobayashi, T. Akagi. Continuous Monitoring of the Methane Concentration in the Atmosphere by IR Spectrometry with 1.66ìm Diode Laser[J]. Analytical Sciences, 2000, (6):1211~1214
[18]郭栓运.差分光谱光纤气体传感器的基本原理[J].应用光学,1989,(6):28~31
[19]刘文琦,牛德芳.光纤甲烷气体传感器的研究[J].仪表技术与传感器,1999,(1):35~36
[20]叶险峰,汤伟中.CH4气体光纤传感器的研究[J].半导体光电,2000,21(3):218~220
[21]阚瑞峰,刘文清等.可调谐二极管激光吸收光谱法测量环境空气中的甲烷含量[J].物理学报,2005,54(4):1927~1930
[22]张雷,尹王保等.基于红外光谱吸收原理的红外瓦斯传感器的实验[J].煤炭学报,2006,31(4):480~483
[23]夏慧,刘文清等.可调谐半导体激光吸收光谱法对高温甲烷的测量研究[J].光学技术,2008,34(2):298~301
[24]童飞.基于BP神经网络的水上交通事故预测及MATLAB实现[D].武汉:武汉理工大学.2005
[25]Saratchandran, P. Dynamic programming approach to optimal weight selection in multilayer neural networks[J]. ieee, 1991, 2(4):465~467
[26]Timo Sorsa, Heikki N. Koivo. Neural networks in process fault diagnosis[J]. ieee, 1991, 21(4):815~825
[27]靳蕃,范俊波,谭永东.神经网络与神经计算机原理应用[M].成都:西南交通大学出版社.1991
[28]徐欣等.基于FPGA的嵌入式系统设计[M].北京:机械工业出版社.2004
[29]M. Marchiesi et al. Linear array architecture implement the Back-Propagation neural network[J]. Proceedings of Second Italian Workshop on Parallel Architectures and Neural Networks, 1990:345~352
[30]Stephen L. Bade, Brad L. Hutchings. FPGA-Based Stochastic Neural Network-Implementation[J]. IEEE Workshop on FPGAs for Custom Computing Machines, Napa, CA, 1994:189~198
[31]David Abramso et al. FPGA Based Implementation of a Hopfield Neural Network for Solving Constraint Satisfaction Problems[J]. Proceedings of 24th Euromicro Conference, 1998, 2(25~27):688~693
[32]Mamun Bin Ibne Reaz et al. FPGA realization of back-propagation for stock market prediction[J]. ieee, 2002,2(18-22):960~964
[33]Yutaka Maeda, Toshiki Tada. FPGA Implementation of a Pulse Density Neural Network With Learning Ability Using Simultaneous Perturbation[J]. IEEE Transaction On Neural Network, 2003, 14(3):688~695
[34]Nan Bu et al. FPGA Implementation of a Probabilistic Neural Network for a Bioelectric Human Interface[J].ieee, 2004, 3(25~28):III-29~32
[35]Hicham Chaoui, Pierre Sicard, Ahmed Lakhsasi. FPGA Implementation of Neural Network Based Adaptive Control of a Flexible Joint with Hard Nonlinearities[J]. ieee, 2006, 4:3118~3123
[36]F. Smach et al. Design of a Neural Networks Classifier for Face Detection[J]. Computer Science, 2006, 2(3):257~260
[37]马小斌.一种集成型神经网络及其基于FPGA的实现[D].广州:华南理工大学.2003
[38]孙永刚.基于FPGA的神经网络硬件实现研究[D].北京:北京科技大学.2004
[39]石凯贺.改善纳米铁酸镧与高分子复合湿敏元件输出特性的神经网络方法及其FPGA实现[D].大连:大连理工大学.2005
[40]董扬.用FPGA实现基于神经网络的旁瓣抑制研究[D].成都:电子科技大学.2005
[41]刘鸣晓.基于FPGA的神经网络数据估计器的设计[D].镇江:江苏大学.2007
[42]李贺.神经网络实时计算关键技术研究[D].合肥:中国科学技术大学.2007
[43]Simon Haykin. Neural Networks: A Comprehensive Foundation[M]. Beijing: Tsinghua University Press. 2001
[44]张青贵.人工神经网络导论[M].北京:中国水利水电出版社.2004
[45]Martin T. Hagan, Howard B. Demuth, Mark H. Beale.神经网络设计[M].北京:机械工业出版社. 2002
[46]Harris D. C., Bertolucci M. C.对称性与光谱学:振动和电子光谱学导论[M].北京:高等教育出版社. 1988
[47]陆婉珍,袁洪福.近代近红外光谱分析技术[M].北京:中国石化出版社.2000
[48]黄宗碧.电力变压器油中微量乙炔故障气体实时遥测光纤传感器的设计理论与实践[D].武汉:华中理工大学.1998
[49]刘瑾.光谱吸收式光纤甲烷气体传感系统的研究[D].秦皇岛:燕山大学.2004
[50]L. A. Kulakova, V. Nonna, A. Sergey. Si-Te Acousto-optic Modulator for Fiber Optic Gas Sensor Based on Midwave InGaAsSb/InAsSbP Diode Laser[J]. Proceedings of SPIE, 1996,2895(6):267~27l
[51]阎卫平,刘文琦,牛德芳.光纤气体传感器的研究[J].传感技术学报,1998:22~25
[52]Dubaniewicz, H. Thomas, E. Joseph. Fiber Optics for Atmospheric Mine Monitoring[J].IEEE Transactions on Industry Applications,1993,29(4):749~754
[53]徐建春.国外矿用光纤沼气传感器及其传输系统[J].煤炭工程师,1998:47~48
[54]F. A. Muhammad. D-shaped Optical Fiber for Chemical Sensing[D]. Glasgow, Scotland: University of strathelyde. 1993
[55]V. Weldon, P. helan, J. Hegarty. H2S and CO2 Gas Sensing Using a 1.57ìm DFB Laser Diode[J]. Proceedings 2nd European Conference on Optical Chemical Sensors and Biosensors, 1994,26(4):19~21
[56]M. Aidaraliev, N. V. Zotova, S. A. Karandashov. Midwave InAsSbP/InGaAsSb Infrared Diode Lasers as a Source for Gas Sensors[J]. Infrared Physics & Technology, 1996,37(1):83~86
[57]W. L. Bares. Tunable Fiber Laser Source for Methane Detection at 1.68μm[J]. Proceedings of SPIE, Chemical, Biochemical and environmental fiber sensors, 1992, 9(9):98~109
[58]I. Klimant. Recent Investigations in Oxygen Sensing[J]. Proceedings 1st European Conference on Optical Chemical Sensors and Biosensors, 1992, (4):12~15
[59]周孟然,李振璧,朱宗玖.分布式光纤传感瓦斯气体系统的研究[J].中国安全科学学报,2007,17(8):167~170
[60]Zhou Mengran, Li Zhenbi, Zhu Zongjiu. Research on fitting the Second-harmonic Spectrum of the Coal Mine Gas with Multiple Regression Method[J]. Proceedings of 2007 8th International Conference on Electronic Measurement & Instruments, 2007, I:195~199
[61]Zhou Mengran, Li Zhenbi, Zhu Zongjiu. Research on the data processing of photoelectric signal surveying for coal mine gas[J], International Symposium on Photoelectronic Detection and Imaging: Technology and Application 2007, 6622:0o1~0o7
[62]周孟然,李振璧,朱宗玖.光纤传感瓦斯吸收池长度对光谱吸收的影响研究[J].中国煤炭,2007,33(10):48~50
[63]陈林.光纤传感在瓦斯检测中的应用研究[D].淮南:安徽理工大学.2007
[64]郭增军.基于光谱吸收光纤甲烷气体传感系统的研究[D].秦皇岛:燕山大学.2002
[65]Zhou Mengran, Li Zhenbi, Zhong Mingyu, He Gang. The intensity calculation of the gas absorption line by multi-line Voigt fitting[J]. Joural of Coal Science&Engineering, 2008, 14(4):534~537
[66]孙圣和,王廷云,徐颖.光纤测量与传感技术[M] .哈尔滨:哈尔滨工业大学出版社.2000
[67]史长胜.基于神经网络的矿井瓦斯涌出量预测[D] .淮南:安徽理工大学.2003
[68]颜晓花.基于GA-BP算法的下石节煤矿215综采面瓦斯涌出量预测[D].西安:西安科技大学.2007
[69]Peter J. Brockwell, Richard A. Davis.时间序列的理论与方法[M].北京:高等教育出版社. 2001
[70]查特菲尔德.时间序列分析导论[M].北京:宇航出版社.1986
[71]Michael Negnevitsky.人工智能智能系统指南[M].北京:机械工业出版社. 2007
[72]贾真.神经网络在煤矿瓦斯预警系统中的应用研究[D].长春:长春工业大学.2007
[73]国家煤矿安全监察局.煤矿安全规程[M].2004
[74]王艳平.基于危险源理论煤矿瓦斯爆炸和涌出风险预警系统及其应用研究[D].重庆:重庆大学.2006
[75]姜雪松等.VHDL设计实例与仿真[M].北京:机械工业出版社.2007
[76]罗苑棠.CPLD/FPGA常用模块与综合系统设计实例精讲[M].北京:电子工业出版社.2007
[77]周润景,图雅,张丽敏.基于QuartusII的FPGA/CPLD数字系统设计实例[M].北京:电子工业出版社.2007
[78]雷伏容.VHDL电路设计[M].北京:清华大学出版社.2006
[79]王辉等.MAX+plusII和QuartusII应用开发技巧[M].北京:机械工业出版社.2007
[80]迈耶-贝斯.数字信号处理的FPGA实现[M].北京:清华大学出版社.2006
[2]Zhou Mengran, Li Zhenbi. Research on coal-mine gas monitoring system controlled by annealing simulating algorithm[J]. Proceedings of SPIE 3rd International Symposium on Advanced Optical Manufacturing and Testing Technologies, 2007, 6723:27-1~27-8
[3]Zhou Mengran, Li Zhenbi, Zhu Zongjiu. Research of Examination Technology of the Second-harmonic in Coal Mine Gas Detector[J]. Proceedings of 2007 8th International Conference on Electronic Measurement & Instruments, 2007, 4:587~591
[4]陈祥光,裴旭东.神经网络技术及其应用[M].北京:中国电力出版社.2003
[5]戴葵.神经网络实现技术[M].长沙:国防科技大学出版社.1998
[6]赵不贿.在系统可编程器件与开发技术[M].北京:机械工业出版社.2004
[7]赵岷.基于FPGA的控制器研究[D].西安:西北工业大学.2002
[8]W. Jin, G. Stewart, B. Culshaw. Absorption Measurement of Methane Gas Using a Broad Light Source and Interferometric Signal Processing[J]. Optics Letters, 1993, 18(16):337~339
[9]H. Inaba, T. Kobayasi, M. Hirama. Optical-Fiber Network System for Air-Pollution Monitoring over a Wide Area by Optical Absorption Method[J]. Electronics Letters, 1979, (23):749~751
[10]J. P. Dakin, C. A. Wade, D. Pinchbeck. A Novel Optical Fiber Methane Sensor [J]. Proc.SPIE, 1987, 734(5):187~190
[11]H. Tai. Long Distance Simultaneous Detection of Methane and Acetylene by Using Diode Lasers Coupled with Optical Fibers[J]. IEEE Photonics Technology Letters, 1992, 4(7): 804~807
[12]F. D Amato, P. Mazzinghi, F. Castagnoli. Methane analyzer based on TDL’s for measurements in the lower stratosphere: design and laboratory tests[J]. Applied Physics, 2002, B 75:195~202
[13]Koichi Ichimura et al. Fiber optic methane gas detection system[J]. Hitachi Cable Review, 2006, 25:11~15
[14]S. Hunsmann et al. Absolute, high resolution water transpiration rate measurements on single plant leaves via tunable diode laser absorption spectroscopy(TDLAS) at 1.37μm[J]. Applied Physics, 2008, B 92:393~401
[15]A. Fried et al. Tunable infrared laser instruments for airborne atmospheric studies[J]. Applied Physics, 2008, B 92:409~417
[16]J. P. Silveira, F. Grasdepot. CH4 Optical Sensor Using a 1.31ìm DFB laser dioxde[J]. Sensors and Actuators, 1995, 25(8):603~606
[17]T. Nakaya, K. Kobayashi, T. Akagi. Continuous Monitoring of the Methane Concentration in the Atmosphere by IR Spectrometry with 1.66ìm Diode Laser[J]. Analytical Sciences, 2000, (6):1211~1214
[18]郭栓运.差分光谱光纤气体传感器的基本原理[J].应用光学,1989,(6):28~31
[19]刘文琦,牛德芳.光纤甲烷气体传感器的研究[J].仪表技术与传感器,1999,(1):35~36
[20]叶险峰,汤伟中.CH4气体光纤传感器的研究[J].半导体光电,2000,21(3):218~220
[21]阚瑞峰,刘文清等.可调谐二极管激光吸收光谱法测量环境空气中的甲烷含量[J].物理学报,2005,54(4):1927~1930
[22]张雷,尹王保等.基于红外光谱吸收原理的红外瓦斯传感器的实验[J].煤炭学报,2006,31(4):480~483
[23]夏慧,刘文清等.可调谐半导体激光吸收光谱法对高温甲烷的测量研究[J].光学技术,2008,34(2):298~301
[24]童飞.基于BP神经网络的水上交通事故预测及MATLAB实现[D].武汉:武汉理工大学.2005
[25]Saratchandran, P. Dynamic programming approach to optimal weight selection in multilayer neural networks[J]. ieee, 1991, 2(4):465~467
[26]Timo Sorsa, Heikki N. Koivo. Neural networks in process fault diagnosis[J]. ieee, 1991, 21(4):815~825
[27]靳蕃,范俊波,谭永东.神经网络与神经计算机原理应用[M].成都:西南交通大学出版社.1991
[28]徐欣等.基于FPGA的嵌入式系统设计[M].北京:机械工业出版社.2004
[29]M. Marchiesi et al. Linear array architecture implement the Back-Propagation neural network[J]. Proceedings of Second Italian Workshop on Parallel Architectures and Neural Networks, 1990:345~352
[30]Stephen L. Bade, Brad L. Hutchings. FPGA-Based Stochastic Neural Network-Implementation[J]. IEEE Workshop on FPGAs for Custom Computing Machines, Napa, CA, 1994:189~198
[31]David Abramso et al. FPGA Based Implementation of a Hopfield Neural Network for Solving Constraint Satisfaction Problems[J]. Proceedings of 24th Euromicro Conference, 1998, 2(25~27):688~693
[32]Mamun Bin Ibne Reaz et al. FPGA realization of back-propagation for stock market prediction[J]. ieee, 2002,2(18-22):960~964
[33]Yutaka Maeda, Toshiki Tada. FPGA Implementation of a Pulse Density Neural Network With Learning Ability Using Simultaneous Perturbation[J]. IEEE Transaction On Neural Network, 2003, 14(3):688~695
[34]Nan Bu et al. FPGA Implementation of a Probabilistic Neural Network for a Bioelectric Human Interface[J].ieee, 2004, 3(25~28):III-29~32
[35]Hicham Chaoui, Pierre Sicard, Ahmed Lakhsasi. FPGA Implementation of Neural Network Based Adaptive Control of a Flexible Joint with Hard Nonlinearities[J]. ieee, 2006, 4:3118~3123
[36]F. Smach et al. Design of a Neural Networks Classifier for Face Detection[J]. Computer Science, 2006, 2(3):257~260
[37]马小斌.一种集成型神经网络及其基于FPGA的实现[D].广州:华南理工大学.2003
[38]孙永刚.基于FPGA的神经网络硬件实现研究[D].北京:北京科技大学.2004
[39]石凯贺.改善纳米铁酸镧与高分子复合湿敏元件输出特性的神经网络方法及其FPGA实现[D].大连:大连理工大学.2005
[40]董扬.用FPGA实现基于神经网络的旁瓣抑制研究[D].成都:电子科技大学.2005
[41]刘鸣晓.基于FPGA的神经网络数据估计器的设计[D].镇江:江苏大学.2007
[42]李贺.神经网络实时计算关键技术研究[D].合肥:中国科学技术大学.2007
[43]Simon Haykin. Neural Networks: A Comprehensive Foundation[M]. Beijing: Tsinghua University Press. 2001
[44]张青贵.人工神经网络导论[M].北京:中国水利水电出版社.2004
[45]Martin T. Hagan, Howard B. Demuth, Mark H. Beale.神经网络设计[M].北京:机械工业出版社. 2002
[46]Harris D. C., Bertolucci M. C.对称性与光谱学:振动和电子光谱学导论[M].北京:高等教育出版社. 1988
[47]陆婉珍,袁洪福.近代近红外光谱分析技术[M].北京:中国石化出版社.2000
[48]黄宗碧.电力变压器油中微量乙炔故障气体实时遥测光纤传感器的设计理论与实践[D].武汉:华中理工大学.1998
[49]刘瑾.光谱吸收式光纤甲烷气体传感系统的研究[D].秦皇岛:燕山大学.2004
[50]L. A. Kulakova, V. Nonna, A. Sergey. Si-Te Acousto-optic Modulator for Fiber Optic Gas Sensor Based on Midwave InGaAsSb/InAsSbP Diode Laser[J]. Proceedings of SPIE, 1996,2895(6):267~27l
[51]阎卫平,刘文琦,牛德芳.光纤气体传感器的研究[J].传感技术学报,1998:22~25
[52]Dubaniewicz, H. Thomas, E. Joseph. Fiber Optics for Atmospheric Mine Monitoring[J].IEEE Transactions on Industry Applications,1993,29(4):749~754
[53]徐建春.国外矿用光纤沼气传感器及其传输系统[J].煤炭工程师,1998:47~48
[54]F. A. Muhammad. D-shaped Optical Fiber for Chemical Sensing[D]. Glasgow, Scotland: University of strathelyde. 1993
[55]V. Weldon, P. helan, J. Hegarty. H2S and CO2 Gas Sensing Using a 1.57ìm DFB Laser Diode[J]. Proceedings 2nd European Conference on Optical Chemical Sensors and Biosensors, 1994,26(4):19~21
[56]M. Aidaraliev, N. V. Zotova, S. A. Karandashov. Midwave InAsSbP/InGaAsSb Infrared Diode Lasers as a Source for Gas Sensors[J]. Infrared Physics & Technology, 1996,37(1):83~86
[57]W. L. Bares. Tunable Fiber Laser Source for Methane Detection at 1.68μm[J]. Proceedings of SPIE, Chemical, Biochemical and environmental fiber sensors, 1992, 9(9):98~109
[58]I. Klimant. Recent Investigations in Oxygen Sensing[J]. Proceedings 1st European Conference on Optical Chemical Sensors and Biosensors, 1992, (4):12~15
[59]周孟然,李振璧,朱宗玖.分布式光纤传感瓦斯气体系统的研究[J].中国安全科学学报,2007,17(8):167~170
[60]Zhou Mengran, Li Zhenbi, Zhu Zongjiu. Research on fitting the Second-harmonic Spectrum of the Coal Mine Gas with Multiple Regression Method[J]. Proceedings of 2007 8th International Conference on Electronic Measurement & Instruments, 2007, I:195~199
[61]Zhou Mengran, Li Zhenbi, Zhu Zongjiu. Research on the data processing of photoelectric signal surveying for coal mine gas[J], International Symposium on Photoelectronic Detection and Imaging: Technology and Application 2007, 6622:0o1~0o7
[62]周孟然,李振璧,朱宗玖.光纤传感瓦斯吸收池长度对光谱吸收的影响研究[J].中国煤炭,2007,33(10):48~50
[63]陈林.光纤传感在瓦斯检测中的应用研究[D].淮南:安徽理工大学.2007
[64]郭增军.基于光谱吸收光纤甲烷气体传感系统的研究[D].秦皇岛:燕山大学.2002
[65]Zhou Mengran, Li Zhenbi, Zhong Mingyu, He Gang. The intensity calculation of the gas absorption line by multi-line Voigt fitting[J]. Joural of Coal Science&Engineering, 2008, 14(4):534~537
[66]孙圣和,王廷云,徐颖.光纤测量与传感技术[M] .哈尔滨:哈尔滨工业大学出版社.2000
[67]史长胜.基于神经网络的矿井瓦斯涌出量预测[D] .淮南:安徽理工大学.2003
[68]颜晓花.基于GA-BP算法的下石节煤矿215综采面瓦斯涌出量预测[D].西安:西安科技大学.2007
[69]Peter J. Brockwell, Richard A. Davis.时间序列的理论与方法[M].北京:高等教育出版社. 2001
[70]查特菲尔德.时间序列分析导论[M].北京:宇航出版社.1986
[71]Michael Negnevitsky.人工智能智能系统指南[M].北京:机械工业出版社. 2007
[72]贾真.神经网络在煤矿瓦斯预警系统中的应用研究[D].长春:长春工业大学.2007
[73]国家煤矿安全监察局.煤矿安全规程[M].2004
[74]王艳平.基于危险源理论煤矿瓦斯爆炸和涌出风险预警系统及其应用研究[D].重庆:重庆大学.2006
[75]姜雪松等.VHDL设计实例与仿真[M].北京:机械工业出版社.2007
[76]罗苑棠.CPLD/FPGA常用模块与综合系统设计实例精讲[M].北京:电子工业出版社.2007
[77]周润景,图雅,张丽敏.基于QuartusII的FPGA/CPLD数字系统设计实例[M].北京:电子工业出版社.2007
[78]雷伏容.VHDL电路设计[M].北京:清华大学出版社.2006
[79]王辉等.MAX+plusII和QuartusII应用开发技巧[M].北京:机械工业出版社.2007
[80]迈耶-贝斯.数字信号处理的FPGA实现[M].北京:清华大学出版社.2006