【作者】 吴贤国；

【导师】 李惠强；

【作者基本信息】 华中科技大学 ， 工程管理， 2006， 博士

【摘要】 工程失败知识管理及预警研究在国际上刚刚开始。本文在国家自然科学基金项目《失败学理论方法与建筑工程失败检测预警研究》(70171049)的支持下,对失败知识管理及工程失败预警进行研究,论文研究工作具有重要的理论意义和实用价值。本文分析了知识管理及其研究现状。探讨了知识管理的过程。对知识管理过程的知识生产、知识共享、知识应用、知识创新进行了分析。探讨了知识管理的过程沿知识螺旋上升的本质。指出建设业知识按知识范畴可分为领域知识、组织知识和项目知识;按知识实践来源,建设业知识可分为最佳实践知识、失败知识和其他知识,其中最佳实践知识和失败知识是建设业知识管理的重点。根据建设业特点,提出建设业知识管理战略为:建设业知识管理的基础是建立行业的学习机制,对最佳实践知识和失败知识的管理是建设业知识管理的两个重点,建设业知识管理的三个角色是建设业行业管理机构、从业团体和IT技术提供者。分析了知识库的构建方法和知识库系统的构成;针对知识表达进行研究,主要探讨了神经网络知识表达方式和应用;指出基于神经网络的专家系统实现了知识表示、存储和推理三者融为一体,具有自组织、自学习和自适应的特点。将神经网络在建筑业知识表达、知识获取中进行应用,采用神经网络对建筑工程工料估算进行知识表达、知识获取应用研究,得到较好的效果。分析了建筑业传统总结经验知识方法的特点,提出了完善的项目学习方法应该具有的特点。建立了基于事件的项目学习方法,此方法以事件的控制过程为核心,通过分析、学习、再工程关键活动将关于事件的信息层逐步引伸到关于事件的知识层,使团队能够在日常的管理活动中不断的积累经验知识,形成一个持续的学习过程。探讨了知识管理系统设计原则、知识管理系统的的特征、知识管理系统的功能、知识管理系统的的组成、知识管理系统的结构、企业知识管理系统的评价体系,并采用层次分析法和专家咨询法对企业知识管理系统评价指标进行了分析。指出我国建筑企业要实施构建适合自己的知识管理系统,应设立知识主管、改进组织结构、建立企业信息化网络、变革企业文化。探讨了工程失败原因及其阶层性,提出了工程项目失败判别指标、工程项目失败影响因素,研究了失败知识化过程和建立失败知识库过程,设计了失败知识库数据表结构和基于PDCA循环知识更新过程。分析了工程失败的监测预警系统,针对模糊神经网络及其在深基坑变形控制中的应用进行研究,进行了深基坑支护变形的模糊神经网络建模与预测,预测结果说明该方法满足工程实际应用精度要求。更多还原

【Abstract】 The engineering failure knowledge management and early warning research has just been started recently. Under the support of National Natural Science Fundation Item: Failure theory method and construction engineering failure examination and early warning research (70171049), the text studies failure knowledge management and engineering failure early warning. This research is of great importance both in theory and in practice.In the text, the author summarized the present research of knowledge management and discussed the process of it, then analyzed the production, sharing, application and innovation of knowledge in the process. The nature of the process is to spiral. The author pointed out that the construction industry knowledge can be divided into domain knowledge, organization knowledge and event knowledge according to the knowledge category; or the best practical knowledge, failure knowledge and other knowledge according to the origin,among which the best practical knowledge and failure knowledge are the key points of constuction industry knowledge management. Considering the characteristics of construction industry, a strategy of constuction industry knowledge management was put forward, with the establishment of the study mechanism being the foundation, the management of the best practical knowledge and failure knowledge being the two key points, the industry govern, the employee group and IT technique promoter being the three participators.The author also analyzed methods of establishing a knowledge base and constitutions of knowledge base system. Researches on expressions of knowledge were carried on, mainly discussing knowledge expression methods for neural network and its application. Based on neural network,the expert system accomplished the purpose that to combine expression,storage with reasoning, and it’s self-organizing, self-learning and self-adapting. We get a good result when using neural network for knowledge expression and possession in construction material estimation.This text analyzed the characters of conventional method of summarizing experience and knowledge of constraction industry, then presumed characters of a perfect learning method. A method was established based on events with controlling process as the key. This method transit from the information layer to the knowledge layer by analyzing, learning and reengineering. It helps to accumulate experience and knowledge gradually in daily management and to form a learning circle.The text studied design principles, characters, functions, structure and evaluation system of knowledge management system, and adopted the layer analysis method to evaluate enterprise knowledge management system. The author suggested that construction enterprises of China establish suitable knowledge management system, they ought to erect a position for knowledge director , improve the structure of the organization,establish an information-based network,and change the corporation culture. In the text,categories of failure and its cause were introduced. Index and influencing factors of engineering failure werw involved. The process of failure experience turning into knowledge and the process of establishing failure knowledge base were discussed. The data form structure of failure knowledge base and renew process of knowledge based on PDCA were designed.The author introduced engineering failure early warning system, studied fuzzy neural network and its application in deep pit deformation control. A fuzzy neural network modle was set up, aiming to forecast deep pit deformation. The result showed that the accuracy of this method may satisfy practical projects.更多还原