Modeling and Automation of Industrial Construction Operations

工业建设项目是加拿大经济的重要组成部分。这些项目涉及能源行业的巨额投资，支持数千个就业岗位，并面临海外建筑公司的竞争。模块化结构是工业建筑的常见做法，特别是在加拿大油砂项目中。在模块化建筑中，大部分项目都是由预制组件组装而成，这些组件是根据独特的设计在现场外定制的。这些组件包括数千个管道线轴、设备和容器以及结构钢元件。它们被单独运送到现场，或预先组装成模块化单元进行最终安装。工业建筑项目通常是快速跟踪的，涉及数百万工时，由几方沿着工程，采购，制造和现场安装阶段提供。这些阶段的运营绩效受到不同组件的独特设计、施工期间发生的设计变更、不同供应商的材料交付偏差以及资源可用性的影响。这导致了一个相当复杂的供应链，如果没有适当的管理和优化，很容易导致项目成本和进度超支。优化这些业务对于保持该部门投资者的最佳价值和保持加拿大承包商的优势，使其能够在竞争日益激烈的全球市场上竞争和出口其专门知识至关重要。 这项研究的目的是调查和开发模型，以代表在关键的工业建设项目交付阶段的操作，并提出替代解决方案，以优化这些阶段的性能。更具体地说，它将集中在1）组件数据结构的分析和表示。这些组件是为每个项目设计的，但在项目之间有一定程度的重复。我们的目标是代表这些组件的数据结构的方式，使我们能够分析项目之间的相似性，从而更好地估计资源需求和进度预期，并开发优化这些组件到模块的包装方法，以尽量减少现场工作。该研究还将针对2）开发新的流程建模方法，适用于捕获现场安装操作的动态，其中通常存在不同步的组件交付，贸易工时可用性的变化以及拥塞限制，并且无法通过传统的规划工具轻松处理。最后，本研究将建议并测试替代方法，以使用选择的数据挖掘算法，在非现场制造和装配操作期间规划和预测交货日期。

Industrial construction projects represent a significant sector of the Canadian economy. These projects involve large investments from the energy sector and support thousands of jobs and are subject to competition from overseas construction companies. Modular construction is a common practice for industrial construction especially on Canadian oil-sand projects. In modular construction, most of the project is assembled from pre-fabricated components that are custom built off-site according to unique designs. These components include thousands of pipe spools, equipment and vessels, and structural steel elements. They are shipped to site either separately or pre-assembled in modular units for final installation. Industrial construction projects are often fast tracked and involve millions of man-hours provided by several parties along the engineering, procurement, fabrication, and site installation phases. The performance of operations in these phases is affected by the unique designs of the different components, design changes that happens during construction, deviations in material delivery by different suppliers, and availability of resources. This results in a fairly complex supply chain that can easily cause project cost and schedule overruns if not properly managed and optimized. Optimizing these operations is of great importance to retain the best value for investors in this sector and to maintain an advantage for Canadian contractors to be able to compete and export their know-how in an increasingly competitive global market. The objective of this proposed research is to investigate and develop models to represent operations in key industrial construction project delivery phases and suggest alternative solutions to optimize the performance during these phases. More specifically it will focus on 1) the analysis and representation of components data structures. These components are uniquely designed for each project but share some degree of repetition between projects. We aim at representing the data structures of these components in a way that allows us to analyze the similarities between projects and thus have better estimates of resource requirements and schedule expectations and develop methods for optimizing the packaging of these components into modules to minimize field work. The research will also target 2) the development of new process modeling approaches suitable for capturing the dynamics of site installation operations where out-of-sync component delivery, variation in trade man-hours availability, and congestion constrains are usually present and cannot be easily handled by traditional planning tools. Finally, the research will 3) suggest and test alternative methods for planning and predicting delivery dates during off-site fabrication and assembly operations using select data mining algorithms.