Information Quality Assessment for Facility Management

Assessing the quality of building information models (BIMs) is an important yet challenging task within the construction industry as projects are increasingly being delivered with BIM. This is particularly essential for facility management (FM) users as downstream information consumers that depend on the quality of models developed in the previous project phases. The research presented in this paper addresses this challenge by introducing a framework for information quality assessment (IQA) of BIMs for FM uses. The IQA framework is the outcome of an extensive study of two large owner organizations involving numerous BIM projects. The framework is structured based on the essential FM subjects: assets, spaces, and systems, and the model characteristics: objects, attributes, relationships, and spatial information. The framework is then operationalized through the development and evaluation of information quality (IQ) tests using BIM model checking tools across three projects with different levels of detail and complexity. The proposed IQA framework and associated tests advance the state of knowledge about BIM quality in terms of methods to represent and evaluate conformance to owner requirements.     

Graph-based version control for asynchronous BIM collaboration

Collaboration and communication are two essential aspects of Building Information Modeling (BIM). Current practice and international standards implement BIM collaboration on the basis of domain model federation where loosely coupled models are managed as separated files and coordinated in a mostly manual fashion. The concept has severe limitations regarding concurrency and version control, as the granularity of change tracking remains on the level of complete files and does not reach individual model objects. Due to this lack of change traceability, high manual effort for the subsequent coordination across the domains is generated. These limitations can be overcome by implementing modern approaches of digital collaboration based on object-level synchronization, widely denoted as BIM level 3. This paper presents a sound methodological basis for object-based version control by (1) representing the object networks of BIM models as formal property graph structures and (2) describing changes of the model by graph transformations. Consequently, modifications can be transmitted as graph transformation rules which are subsequently integrated on the receiving side, thus achieving object-level synchronization. The paper provides the underlying theory of describing model changes by means of graph transformations and demonstrates its benefits using the example of domain models implementing the Industry Foundation Classes (IFC) as their underlying data model.     

SpaRSE-BIM: Classification of IFC-based geometry via sparse convolutional neural networks

Information exchange between Building Information Modeling (BIM) tools is challenging, since many applications use their own native data formats. The Industry Foundation Classes (IFC) schema, an open data exchange format for BIM, does not capture the full semantic meaning needed for direct use by different BIM tools and can be prone to information loss due to reduction, simplification, translation and interpretation of the data. Current practice often treats the imported model as a reference and requires a user to remodel the building using the respective application’s native elements. Many BIM object properties are defined by its classification. Inconsistencies in the mapping between native BIM elements and IFC, e.g. due to unsupported export functionality or manual error, can lead to problems when using the model in a downstream application. Recent works demonstrate that neural networks offer a promising possibility to alleviate this issue via classification of the objects contained in a BIM model and suggesting those corrections to the user. However, the computational overhead of these deep learning models, either due to necessary pre-processing of the data or runtime performance of the model, makes it difficult for them to be used in plug-ins or middleware for BIM tools. This work proposes SpaRSE-BIM, a neural network model based on sparse convolutions for the classification of IFC-based geometry and semantic enrichment of BIM models. Experiments are performed on two IFC entity classification benchmark datasets. The results demonstrate that SpaRSE-BIM is significantly more efficient at inference time compared to previous approaches, while maintaining state-of-the-art accuracy. Further experiments explore the applicability of IFC entity classification datasets to the domain of Scan-to-BIM. It can be shown that the feature space of SpaRSE-BIM learns to discern objects in a semantically meaningful way, even in cases where fine-grained subtype information for IFC objects is not available during training.     

Analysis of early-design timber models for sound insulation

Timber construction is associated with a low carbon footprint and offers a high degree of sustainability. However, it poses challenges considering sound insulation. Acoustic analyses, which could require major expensive and time consuming changes in the building design, are typically performed once the design is already in the detailed stage. By using building information modelling (BIM), it is possible to shift the planning of the building physics, including acoustic analysis, to earlier phases. To make this possible, building models must include all the information necessary to perform acoustic analyses. One important part of acoustic analysis is identifying junctions between elements and map them to the junction types in standards. Until now, this investigation involves tedious manual processing for extracting multiple topological dependencies between different elements. Hence, this paper presents a framework for a seamless workflow between building models and acoustic analysis tools, based on an analysis of data models. The framework extracts and analyzes the element types, their geometry, and the connections of the individual elements in relation to each other. Through topological reasoning, along with a set of logical rules, the proposed framework identifies fifteen types of junctions, which can be distinguished acoustically for timber construction. The approach was evaluated in a prototypical implementation using a real-world model based on Industry Foundation Classes (IFC) as an example, in which the potential connection types were successfully extracted. This paper shows that junction analysis can be done with a geometric analysis to fill in missing semantic information about junctions of elements from the original data model.     

Developing final as-built BIM model management system for owners during project closeout: A case study

To apply final as-built BIM models to facility management (FM) during the operation phase, it is important for owners to obtain an accurate, final as-built model from the general contractors (GCs) following project closeout. Confirming the accuracy of the final as-built BIM model is one of the most important works executed by owners to meet the accuracy requirement of final as-built models for FM. However, many practical problems arise relating to the management of final as-built models such as final as-built model mismatch, the lack of available final as-built models, and the entry of incorrect non-geometric information into the final as-built models. To solve these practical problems, this study develops a Final As-built BIM Model Management (FABMM) system for owners to handle final as-built BIM model inspection, modification, and confirmation (BMIMC) work beyond project closeout. The proposed approach and system can be used to manage the status and results of BMIMC management work for the final as-built BIM model to be performed. The proposed approach and system were applied in a case study in a selected building in Taiwan to verify and demonstrate its practical effectiveness. This study identifies the benefits, limitations, and conclusions of the FABMM system, and presents suggestions for its further application.     

BIM 模型相似度计算方法

针对从业者不论是想从本地模型库还是线上共享网站获取所需的BIM 模型只能靠 逐个查找、人工识读的方法,而模型的数量越来越多,获取符合需求的模型需要花费大量的时 间和人力的问题,提出了一种构件级BIM 模型相似度计算方法。从模型的构件出发,以BIM 通用交互格式工业基础类(IFC)文件作数据源,以通用数据标准IFC 2×3 为数据基础,首先提取 模型中构件的几何信息、语义信息等,并利用改进的方向包围盒(OBB)碰撞检测算法查找相连 构件;然后以构件为顶点、构件间连接关系为边将BIM 模型构建为邻接图模型,并用图编辑距 离算法计算邻接图模型的编辑距离;最后即可计算出不同模型之间的相似度。该方法以构件级 BIM 模型的相似度为依据可以大大提升BIM 模型的检索速度与准确率。     

Automated facility inspection using robotics and BIM: A knowledge-driven approach

Facility inspection is crucial for ensuring the performance of built assets. A traditional inspection, characterized by humans’ physical presence, is laborious, time-consuming, and becomes difficult to implement because of travel restrictions amid the pandemic. This laborious practice can potentially be automated by emerging smart technologies such as robotics and building information model (BIM). However, little has been known on how such automation can be achieved, concerning the knowledge-intensive nature of facility inspection. To fill the gap, this research aims to develop a knowledge-driven approach that can synergize knowledge of diverse sources (e.g., explicit knowledge from BIM, and tacit experience in human minds) to allow autonomous implementation of facility inspection by robotic agents. At the core the approach is an integrated scene-task-agent (iSTA) model that formalizes engineering priori in facility management and integrates the rich contextual information from BIM. Experiments demonstrated the applicability of the approach, which can endow robots with autonomy and knowledge to navigate the challenging built environments and deliver facility inspection outcomes. The iSTA model is publicized online, in hope of further extension by the research community and practical deployment to enable automated facility inspection using robotics and BIM.     

BIM 空间关系数据的云存储与检索方法研究

随着工程项目体量的增大,模型中存在的空间关系更加复杂多样化,现有的建筑 信息模型(BIM)数据存储和检索方式无法满足使用要求。为了提高 BIM 中大量空间关系数据的 存储和检索效率,通过集成 BIM 与云计算技术,提出了一种 BIM 分布式负载均衡集群方案, 在此基础上利用弹性搜索框架(Elastic Search)和图数据库Neo4j实现了IFC空间关系数据的云存 储和检索,为海量 BIM 空间关系数据提供一种高效快速的云存储和检索方法。     

An approach to combine progressively captured point clouds for BIM update

Building information models (BIMs) provide opportunities to serve as an information repository to store and deliver as-built information. Since a building is not always constructed exactly as the design information specifies, there will be discrepancies between a BIM created in the design phase (called as-designed BIM) and the as-built conditions. Point clouds captured by laser scans can be used as a reference to update an as-designed BIM into an as-built BIM (i.e., the BIM that captures the as-built information). Occlusions and construction progress prevent a laser scan performed at a single point in time to capture a complete view of building components. Progressively scanning a building during the construction phase and combining the progressively captured point cloud data together can provide the geometric information missing in the point cloud data captured previously. However, combining all point cloud data will result in large file sizes and might not always guarantee additional building component information. This paper provides the details of an approach developed to help engineers decide on which progressively captured point cloud data to combine in order to get more geometric information and eliminate large file sizes due to redundant point clouds.     

输变电设备运维知识图谱的构建及应用

针对电网公司在开展输变电设备运维检修过程中存在的非结构化文本数据难以利用、全口径数据难以深度融合、数据应用手段仍处于简单统计状态等难题,本文研究设计了基于知识图谱的输变电设备运维知识问答系统的总体框架,以打破传统信息获取方式的局限性。首先,从多源输变电设备数据中提取出实体、关系、属性等元素,涵盖的数据既有结构化也有非结构化的数据。在此过程中,使用不同的技术和方法来提取这些元素。其次,利用图形数据库Neo4j将这些元素组合在一起,构建一个完整的输电变电设备运维知识图谱,从而将不同类型和业务数据整合并形成一个连贯统一的知识网络。最后,基于Neo4j +Python平台搭建输变电设备运维知识问答系统,为运维人员提供一种新的方法。     

宠物知识图谱的半自动化构建方法

提出一种宠物知识图谱的构建框架。通过自顶向下的方式设计并构建了schema（概念）层,从半结构化和非结构化数据中进行知识抽取构建了数据层。在对非结构化数据的实体抽取方面,提出了一种条件随机场（CRF）与宠物症状词典相结合的症状命名实体识别方法。该方法利用症状词典对文本进行识别,获取语义类别信息,CRF结合语义信息实现对症状实体的识别抽取。实验结果表明了该方法的有效性。在知识表示方面,选用OrientDB数据库支持的属性图模型来表示。知识图谱采用OrientDB图数据库来完成知识的存储,并实例展示了构建的宠物知识图谱。     

A relational framework for smart information delivery manual (IDM) specifications

This study proposes a relational framework for standardized machine-applicable, readable, and transferable (smart) information delivery manual (IDM) specifications. As the demand for projects using building information modeling (BIM) increases, there is a more acute awareness and need for clearly defined information requirements to support the BIM projects. The ISO 29481-1 IDM standard defines how to specify exchange requirements (ERs) and their use cases (UCs) using a process map (PM). However, IDM specifications are currently not easily sharable or reusable due to the lack of a commonly accepted standard data schema. This study overcomes this problem by identifying definitions and relationships of IDM components to develop the relational framework for an IDM data schema. An extensive review of existing IDM-related documents and standards, and iterative international meetings, were conducted by 46 international IDM experts from 16 countries. A consensus on the relational framework was then reached through three Delphi survey rounds. The formalized relational IDM framework lays a foundation for further developing an IDM data schema as an international standard.     

基于BIM技术的建筑建模与成本控制

本论文根据潍坊市寒亭区农村住宅F9C住宅楼,建筑面积195.5 m2,使用面积169.7 m2,根据住宅图纸建立模型、通过建筑建模解决问题、运用BIM建模发现的问题。建立模型当中屋顶的绘制和玻璃斜板的绘制是难点,出于图纸设计和整体美观的考虑,改变部分墙体、楼板和玻璃斜板的坡度,使屋顶更加美观,优化了玻璃斜板的施工方法。由于BIM建模处于施工阶段的前一项工作,能够全面地体现整体建筑,做好数据信息和几何信息的控制,同时可以监控施工中工程信息的变化的同时更改信息实时记录,弥补了传统成本控制中事前控制和事中控制的缺失,不仅对图纸设计进行改良避免了施工过程的损失,还对工期和质量的控制保障有所提高。对于BIM这个新兴产品,在国内没有被广泛的接受,原因不单单包括大部分人对BIM的陌生和只有少数人拥有BIM技术,还包括计算机技术的提升和投资方不愿意对该业务资金的投入,BIM在国内发展缓慢。经过时间累计,BIM对工程成本控制的优点会被越来越多的人看到,也将会带动建筑业的发展。     

BIM 技术在多灾害领域中的应用研究综述

建筑信息模型(BIM)是建筑行业的热门技术，多灾害是建筑防灾减灾领域的前沿问 题。将 BIM 技术应用在多灾害研究，可以为多灾害研究提供共享、统一的数据模型，既提高了 信息利用率，又保证不同灾害分析的模型一致性。本文通过中国知网和 Web of Science 对国内 和国际上 BIM 与地震、风灾和火灾的研究进行分析，总结了各个方向的主要关键词。整体上， BIM 与多灾害的研究处于起步阶段，有待于进一步发展。     

A BIM-Oriented Model for supporting indoor navigation requirements

Existing indoor navigation approaches such as navigation based on 2D geometries and pre-defined routing remain insufficient for many applications such as emergency response, delivery, utility maintenance and facility management. The insufficiencies caused by existing navigation approaches can be overcome by making use of the advanced semantic and geometric information included in intelligent building models. A key example of such models is Building Information Models (BIMs) which contain detailed geometric and semantic information about buildings. In fact, the BIMs’ structure is very complex for facilitating navigation. This paper presents a new BIM Oriented Modeling methodology resulting in the definition of a new BIM based model (BO-IDM) dedicated for facilitating indoor navigation. The paper later describes the transformation of information from a standard BIM (IFC) into the new model (BO-IDM). The innovation aspects of BO-IDM can be summarized as follows: (i) it provides highly detailed semantic information for indoor navigation and (ii) it represents the non-geo-referenced structure and complex geometries of BIMs with ISO 19107 compliant representations. Therefore this model is well suited for indoor navigation.     

基于BIM与知识图谱的智能化审图系统设计与实现

针对传统建筑审图存在的人工依赖性强、数据难融合共享和审查内容可视化水平低等问题,提出一种基于BIM与知识图谱的建筑智能审图方法;首先,利用知识图谱技术将繁杂的建筑规范文本数据关联为结构化知识,建立知识与数据间的语义联系;其次,鉴于BIM模型的多源异构特性,提出基于IFC的建筑数据信息提取整合,为后续建筑智能化审图实现提供数据基础;最后基于HTML5网络框架和人工智能技术设计开发了具有三维可视化、跨平台等功能的建筑智能审图系统,实现了建筑模型网页端显示及可交互操作、审图内容可视化展示、规范条文对应显示等功能;经实际建筑项目实验测试,该系统有效实现了被审建筑模型的智能化审图,并自动检测出不符合规范的建筑模型构件,一定程度上提升建筑行业智能化审图水平。     

Detecting logical relationships in mechanical,electrical, and plumbing (MEP) systems with BIM using graph matching

The increasingly complex design has gained difficulty in conducting the rule compliance checking for the Mechanical, Electrical and Plumbing (MEP) system in the design phase. Useful rule-checking systems could contribute to a quicker project delivery time. Currently, an efficient method for checking the logical relationship is still lacking. This study aims to propose an MEP rule checking framework using the subgraph matching technology. First, the MEP components in the BIM model are extracted by utilizing the application programming interface (API), and a graph database is established with point-based and curve-based instances being nodes and relationships, respectively. Second, the graph database is simplified to increase the speed of graph matching. Third, the rules, which regulate how the MEP components should be connected, are represented by a knowledge graph. Finally, rule checking is achieved by comparing the graph database against the knowledge graph, and the critical path in a sub-system is detected by calculating the betweenness centrality. A case study with a rail station is used to evaluate the approach where the overall model checking and rule checking are conducted on the original and simplified graph databases sequentially. The results show that the proposed approach could achieve the rule compliance checking at a high speed, and 6 unconnected instances along with 155 problematic pipe fittings have been found. Besides, the critical path for the selected ACS system is from the water-cooled chiller to the condenser water pump. The proposed framework could help in the overall model checking and rule checking process, improving the efficiency of BIM engineers. This research demonstrates that converting a BIM model into a graph database can benefit conventional BIM analysis methods by incorporating advanced technologies (e.g., artificial intelligence) to enable a more flexible and accurate MEP design process.     

城市大脑知识图谱构建及应用研究

随着城市大脑建设进程的推进,城市中积累了大量的物联网(IoT)设备和数据,利用海量设备数据对问题进行分析和溯源,对于城市大脑建设具有重要意义。该文基于资源描述框架和智能物联网协议概念,提出一种以城市物联网本体为基础的城市大脑知识图谱建设方法,城市大脑知识图谱模型融合多源异构数据,覆盖城市基本要素,实现对城市要素的全面感知和深度认知。该文重点探究了城市事件本体中的事件抽取,设计了一种新颖的语言模型框架对事件类型和论元联合抽取,与单模型分析对比,该联合模型较单模型的事件类型和论元F₁值分别提高0.4%和2.7%,在时间和模型复杂度上,较单模型级联也有更好效果。最后,该研究对知识图谱技术与人工智能、多传感器融合、GIS等新一代信息技术交叉融合方面进行了探究分析,为城市治理和服务应用场景提供理论依据。     

融合实体描述信息和邻居节点特征的知识表示学习方法

知识图谱表示学习旨在将实体和关系映射到一个低维稠密的向量空间中。现有的大多数相关模型更注重于学习三元组的结构特征,忽略了三元组内的实体关系的语义信息特征和三元组外的实体描述信息特征,因此知识表达能力较差。针对以上问题,提出了一种融合多源信息的知识表示学习模型BAGAT。首先,结合知识图谱特征来构造三元组实体目标节点和邻居节点,并使用图注意力网络（GAT）聚合三元组结构的语义信息表示;然后,使用BERT词向量模型对实体描述信息进行嵌入表示;最后,将两种表示方法映射到同一个向量空间中进行联合知识表示学习。实验结果表明,BAGAT性能较其他模型有较大提升,在公共数据集FB15K-237链接预测任务的Hits@1与Hits@10指标上,与翻译模型TransE相比分别提升了25.9个百分点和22.0个百分点,与图神经网络模型KBGAT相比分别提升了1.8个百分点和3.5个百分点。可见,融合实体描述信息和三元组结构语义信息的多源信息表示方法可以获得更强的表示学习能力。     

结合五笔字形与上下文相关字向量的命名实体识别

命名实体识别（NER）作为自然语言处理的重要部分,在信息抽取和知识图谱等任务中得到广泛应用。然而目前中文预训练语言模型通常仅对上下文中的字符进行建模,忽略了中文字符的字形结构。提出2种结合五笔字形的上下文相关字向量表示方法,以增强字向量的语义表达能力。第一种方法分别对字符和字形抽取特征并联合建模得到字向量表示,第二种方法将五笔字形作为辅助信息拼接到字向量中,训练一个基于字符和五笔字形的混合语言模型。实验结果表明,所提两种方法可以有效提升中文NER系统的性能,且结合五笔字形的上下文相关字向量表示方法的系统性能优于基于单一字符的语言模型。