【作者】 孔令东；

【导师】 张虹；

【作者基本信息】 中国矿业大学 ， 地图制图学与地理信息工程， 2011， 博士

【摘要】 煤与瓦斯突出是严重影响煤矿安全生产的自然灾害之一,依据瓦斯突出机理因素、因素权重和时空序列,引入模糊Petri网和工作流挖掘技术,描述瓦斯突出的动态演变过程,进而用于瓦斯突出预测的推理分析,可为突出预测研究、突出前兆动态信息的处理及应用提供一种新的理论方法和技术途径。本文在对瓦斯突出预测方法和模糊Petri网理论深入研究的基础上,实地调研分析瓦斯突出的空间分布,提出模糊Petri网的瓦斯突出预测模型,分别从瓦斯突出的模糊Petri空间建模、体系结构、模糊空间Petri网的规则描述和动态推理、模型系统的实现等几个方面进行深入和系统的研究。提出了瓦斯突出的模糊Petri网空间模型,针对瓦斯突出各因素空间分布的特点,解决了瓦斯突出的模糊性描述和空间建模问题。在深入分析Petri网和模糊Petri网相关定义、模型结构、空间关联动态性质和模糊知识表示的基础上,研究了模糊Petri网和瓦斯突出的空间预测的结合点；深入研究瓦斯突出的模糊空间表示,扩展模糊Petri网的相关组件成分,建立瓦斯突出的模糊Petri网空间模型,使瓦斯突出影响因素在不同的空间位置通过空间模型表示出来。提出了基于工作流网的瓦斯突出空间数据挖掘,实现了瓦斯突出空间状态因素和业务流程建模的统一,解决瓦斯突出空间模型的可实现性和空间决策问题。在论证空间数据挖掘和工作流结合的必要性和可行性基础上,分析工作流及过程优化,研究了基于Petri网的工作流网,从定义、组件和Petri网进行工作流建模的优点等几方面进行了对比分析；通过扩展工作流参考模型的接口,提出了基于工作流网的空间数据挖掘体系结构,进一步结合瓦斯突出的空间预测研究,建立瓦斯突出工作流网模型,将瓦斯突出空间预测模型作为工作流过程定义模型,挖掘瓦斯突出预测的规则,用于瓦斯突出预测的过程；从静态数据挖掘和动态数据额挖掘两个方面,在对关键技术实现研究的基础上,完成了实例分析,突破了模糊Petri网的瓦斯突出模型的可实现性瓶颈,建立了模型系统的体系结构。提出了模糊空间Petri网,为增强模糊Petri网空间描述的通用性,将空间属性作为重要指标,解决了瓦斯突出不同空间状态建模的动态关联和推理等可实现性问题,从定义、组件、知识表示、规则描述和动态推理等方面形成了较为完整的空间建模理论体系。深入研究空间位置和模糊Petri网的结合,引入空间关联因子,借鉴相关扩展的模糊Petri网,给出了模糊空间Petri网的定义；依据空间影响因子分布情况,详细地阐述了状态关联影响组件、变迁关联影响组件、状态关联多值阈值的构成及含义,解决了瓦斯突出不同空间区域关联的知识表示问题；随后,对比基本Petri网结构进行了等价性分析,对比基本模糊Petri网进行了改进性分析；深入地研究了模糊空间Petri网的知识表示,详细描述了相关分层的思想、面向对象表示和XML表示；最后,详细研究了动态推理和模糊空间Petri网规则描述,提出了一种带反馈的动态推理算法。深入探索Petri网的可实现性,建立基于WF的瓦斯突出预测模型系统,解决了瓦斯突出Petri网建模的程序可实现性。基于微软最新工作流开发平台WF4.0,实现了基本Petri网和模糊空间Petri网的建模和组件开发。在此基础上建立了模型验证系统,可以描述模糊Petri网的规则推理和动态测试。该论文有图72幅,表10个,参考文献157篇。更多还原

【Abstract】 Coal and gas outburst presented obviously space-time and multidimensional complexity, describing dynamic evolvement process was used to infer and analyze the gas outburst according to gas outburst mechanism factor, factor weight and space-time sequence. Fuzzy Petri net spatial forecast modeling of gas outburst was presented, which thoroughly and systemically researched fuzzy Petri net model, architecture, rule description and dynamic inference of fuzzy spatial Petri net, realization of model system.Fuzzy Petri net spatial model of gas outburst was presented according to space distribution character of outburst diversified factor, and fuzzy description and spatial model of gas outburst were solved. The integration of fuzzy Petri net and gas outburst spatial forecast was researched. The component of fuzzy Petri net was extended integrating with space distribution character of outburst diversified factor. Lucubrating Fuzzy spatial representation of gas outburst, effect factor of gas outburst was represented by spatial model in the different spatial place.Spatial data mining of gas outburst based on workflow net was presented in order to realize the unification between spatial state factor and business flow modeling of gas outburst and to solve spatial model realization and spatial decision question. Workflow net based on Petri net was researched and spatial data mining architecture based on workflow net was presented by extending workflow reference model. Workflow net model of gas outburst was built combing with spatial forecast of gas outburst and key technology realization was analyzed from static and dynamic data mining.Fuzzy spatial Petri net was presented convenient for describing spatial dynamic association and dynamic reasoning and enhancing spatial describing universality of fuzzy Petri net. Gas outburst knowledge representation of different spatial distribution association was solved according to spatial effect factor distribution, which structures and meanings of state association effect component, transition association effect component and state association multi-valued threshold were illuminated in detailed. Knowledge representation of fuzzy spatial Petri net was researched in detailed and a dynamic reasoning algorithm with feedback was presented.Model System of gas outburst based on WF was built based on probing deeply into Petri net realizability, and program realizability of Petri net modeling for gas outburst was solved. Modeling and developing component of basic Petri net and fuzzy spatial Petri net were implemented based on Microsoft workflow foundation 4.0, and model verification system was built which could described rule reasoning and dynamic testing of fuzzy Petri net.更多还原