装配式建筑施工安全风险涌现机制与动态智能诊控方法研究

For the prefabricated construction (PC), it is the key problem that mining construction risk emergence character and inner law, establishing dynamic intelligent risk diagnoses and control mechanism by PC construction safety risk information under dynamic environment for insuring construction security and healthy development of PC. The emergence mechanism and complex coupling interplay feature of multilevel PC construction safety risk factors are taken as the research object, combing emergence driven case-based reasoning (E-CBR) and artificial immune algorithm (AIA), focusing on inner mechanism for off-line inducing rule learning and on-line inducing scheme reasoning, the research work will be conducted. The detail contents including heteromorphism emergence mechanism of multilevel PC construction safety risk factors; real time and precision driven E-CBR and AIA inner integration mechanism; inducing rule generating method on both positive direction(from diagnoses and control mechanism analysis to rule abstracting beforehand) and inverted direction (from construction risk data to rule mining afterwards); double inducing rule fusion by AIA, multi-risk diagnoses and control rule cooperating strategies and methods; real time diagnoses and control scheme reasoning oriented E-CBR implementing mechanism; integrated optimization method of customization diagnoses and control. By the research work, the complex emergence law of multilevel complex PC construction risk factor will be discovered; the contradiction between precision inducing and real time responses will be eliminated; novel theory and method from risk customization and cooperating, statistic and dynamic integrating dynamic diagnoses and control mode and so on will be produced. Above research work will provides supporting for improving construction safety supervision and control effectiveness, preventing construction accident and risk in PC.

针对装配式建筑(PC)，利用动态环境下施工安全风险信息，挖掘风险涌现特征与内在规律，建立动态智能诊控方法，是确保PC施工安全与健康发展的关键。本研究以PC施工安全风险因素多级涌现、复杂耦合作用为研究对象，融合涌现驱动型案例推理(E-CBR)与人工免疫算法(AIA)，围绕风险诊控离线规则学习与在线方案推理内在机制展开研究。具体包括：异态多级PC施工安全风险涌现机制；实时与精确双驱动的E-CBR与AIA内在集约机制；正向(事前诊控机理分析到规则提取)与逆向(事后风险数据到规则挖掘)兼顾的诊控规则生成方法；基于AIA的双向诊控规则融合、多维诊控规则协同策略与方法；面向即时诊控方案的E-CBR实现机制；定制化诊控方案集成优化。本研究将揭示PC施工安全风险各级因素复杂涌现规律，消除精确诊控与实时反应间矛盾；从定制与协同、静态与动态集成诊控等方面形成新理论方法，为提高PC施工安全风险监管效力提供支持。

在动态环境下，揭示施工安全风险涌现规律，建立动态智能诊控机制是确保装配式建筑(PC)施工安全生产的关键所在。本项目总体上围绕PC施工安全风险涌现与动态诊控问题，研究了PC施工安全风险涌现机制、离线诊控规则学习与在线诊控方案推理三大体系内容。具体揭示了以施工安全“风险元素—风险元组—风险特征—风险状态”为主线的异态多级PC施工安全风险涌现机制；构建了多形态施工安全风险涌现的动力学模型；研究了PC施工安全风险元组、风险特征、风险状态的正向诊控机理与规则化显性表达方法；实现了基于AIA的逆向诊控规则挖掘；探究了正向机理规则与逆向挖掘规则的融合方式；研究了AIA中的抗原(抗体)编码、抗原识别、初始抗体群产生、亲和力计算、克隆选择、群体更新等环节在PC施工安全风险诊控规则参数优化等方面的具体实现方式；实现了面向PC施工安全风险诊控方案动态链接结构优化；构建了定制化的事故概率导向类、事故损失导向类、安全水平导向类、投入水平导向类、其他控制导向类等施工安全风险控制优化模型，给出了求解方法；结合规则的案例化集成，研究了E-CBR在PC施工安全风险诊控方案的涌现型案例表达、涌现型案例匹配、涌现型案例调整与涌现型案例维护技术；实现了实时性与精确性双驱动的E-CBR与AIA集约诊控机制。本项目实现了不同模式下PC施工安全风险涌现计量；形成了“双向”融合的PC施工安全风险诊控规则，挖掘出风险诊控规则的关键参数阈值；优化了PC施工安全风险控制变量数值。本项目克服了PC施工安全风险动态信息获取瓶颈问题；拓展了PC施工安全监管的体系范畴；揭示了PC施工安全风险诊控的多级涌现机制；丰富了PC施工安全风险诊控的角度与模式；优化了PC施工安全风险控制方案；实现了PC施工安全风险诊控机制创新与性能提升。