The current issue and full text archive of this journal is available on Emerald Insight at:

www.emeraldinsight.com/0263-2772.htm

F
36,13/14 Lean-Agile FM-BIM: a
demonstrated approach
J. J. McArthur and Brandon Bortoluzzi
Department of Architectural Science, Ryerson University, Toronto, Canada
676
Received 20 April 2017
Revised 18 September 2017
Abstract
14 February 2018 Purpose – This paper aims to respond to the high cost of facility management-enabled building information
Accepted 16 April 2018 model (FM-BIM) creation and maintenance, a significant and under-researched barrier to adoption for existing
buildings. The resultant approach focuses on only value-adding content (“Lean”) developed flexibly and
iteratively in collaboration with end-users (“Agile”).
Design/methodology/approach – Five case studies were developed for university and hospital
buildings in collaboration with end-users, guided by the process presented. These informed the refinement of
a robust and flexible approach to increase BIM functionality with minimal geometry, focusing instead on the
development of specific parameters to map semantic information necessary for each desired FM use.
Findings – The resulting BIM provided a breadth of model functionality with minimal modeling effort: 15
hours average implementation time per supported FM use. This low level of effort was achieved by limiting
geometry to where it is necessary for the FM use implementation. Instead, the model incorporated the
majority of geometry by reference and focused on semantic and topological parameters to house FM
information.
Research limitations/implications – This study provides the basis for a new ontology structure
focused on defining the rules for hosting asset management data (host entity, parameter type and
characteristics) to reduce the reliance on complex geometric model development.
Practical implications – By prioritizing highly beneficial applications, early investment is minimized,
providing quick returns at low risk, demonstrating the value of FM-BIM to end-users.
Originality/value – The Lean-Agile approach addresses the known research gap of low-effort, flexible
approaches to FM-BIM model creation and maintenance and its effectiveness is analyzed through five case studies.
Keywords Facilities management Lean, Agile, Information management,
Building information modeling (BIM), Data transfer
Paper type Research paper

1. Introduction
Digital technologies have fundamentally transformed our interaction with the built
environment. Mobile devices provide a means to rapidly access and share information.
Within the facilities context, these technologies impact both capital project delivery and day-
to-day operations. Computer-aided facilities management (CAFM) systems supporting the
full range of hard- and soft- facilities management (FM) activities have become ubiquitous.
Enhanced by sensor networks and Internet of Things (IoT) devices, FM teams have access
to a breadth of building information in real time. Mobile apps and cloud-hosted file systems
further enhance this functionality, providing service engineers with field access to building
and equipment information and building occupants to a limited range of self-service
activities such as real-time room scheduling and issue reporting.
Facilities
Vol. 36 No. 13/14, 2018
pp. 676-695 The authors acknowledge financial support received from the Ontario Research Fund (BRAIN
© Emerald Publishing Limited Alliance), the Ryerson Work-Study Research Assistant grant and the participation of Ryerson
0263-2772
DOI 10.1108/F-04-2017-0045 University Department of Facilities Management and Development.
 The design and construction of buildings has been similarly affected by digital Lean-Agile
technologies. Parametric design tools, building information models (BIMs), digital FM-BIM
fabrication and virtual construction scheduling generate a wealth of digital data on the built
environment. When integrated with the CAFM system, this data can provide significant
benefits such as utility cost reductions, comfort management, space optimization, improved
inventory management and energy simulation and conservation (Love et al., 2014). When
used for design and construction, BIMs act as central databases for building information
(Eastman et al., 2011), enriched with semantic information associated with common end 677
uses, summarized by Kreider et al. (2010). FM-enabled BIMs (FM-BIMs) are of specific value
in this digital context, having demonstrated time- and cost-savings benefits for soft (Arayici
et al., 2012) and hard facilities management activities (Volk et al., 2014; Love et al., 2014;
Arayici et al., 2012; Bryde et al., 2013).
Despite the known benefits of FM-enabled BIM (FM-BIM), significant barriers to
adoption exist, particularly for existing buildings where as-built geometry is typically
unavailable (Volk et al., 2014; Gheisari and Irizarry, 2016). Most research focused on
bridging BIM and building is concerned with construction stage tracking and generation of
an as-built model with limited research at other project stages, demonstrating this as a
known research gap (Chen et al., 2015). Similarly, a limited number of FM-BIM case studies
have been published (Kassem et al., 2015). This study provides a significant contribution by
presenting a Lean and Agile approach for the development of FM-BIM for existing buildings
and demonstrating its application in the field. The practical application of this research is
the rapid and reduced-cost deployment of beneficial FM applications, even in buildings with
no pre-existing BIM.

2. The case for a Lean approach to building information models for facility
management
Traditionally, FM-BIMs are built using as-built construction models, to which FM
information is included, ideally using Construction Operations Building Information
Exchange (COBie; East, 2007). Similarly, BIM use by contractors is increasing, with a
majority now using BIM across several trades (Hanna et al., 2013).

2.1 The value of facility management-based building information model

FM-BIM provides operational benefits such as utility cost reductions, comfort management,
space optimization and improved inventory management (Love et al., 2014), and supports
enhanced FM tasks through visualization and analysis capabilities (Becerik-Gerber et al.,
2012). As a result, there has been significant research in developing FM-BIMs for new
buildings (e.g. Kivits and Furneaux, 2013; Kiviniemi and Codinhoto, 2014; Brooks and
Lucas, 2014). A BIM with highly detailed information integrated with shop drawings and
operation and maintenance (O&M) manuals is an oft-stated goal of the construction phase
(Volk et al., 2014; Eastman, et al., 2011) to provide the necessary inputs to support facilities
management activities.

2.2 Building information models-facility management in the existing building context

Despite its benefits, several barriers have prevented the widespread adoption of BIM in
facilities management, with the result that fewer than 10 per cent of BIMS are used during
operations/maintenance (Eadie et al., 2013). Case studies of FM-BIM for existing facilities are
severely limited but have drawn the same conclusions. Kassem et al. (2015) modelled 32
university buildings totaling over 120,000 m2, noting key challenges arising, including
identifying the necessary information – “what information is to be provided, when and by
F whom” – and addressing interoperability issues. In their review, Ilter and Ergen (2015)
36,13/14 similarly note the development of models (building survey and generating as-built BIM),
information exchange, interoperability, integration and access on BIM among the key
challenges for BIM for existing building maintenance and refurbishment. Parsanezhad and
Dimyadi (2013) address these issues with bespoke application developed by the FM team at
Unitec, providing valuable insight on how end-user input can help address these barriers.
678 This latter case study is context-specific, however, and the research gap to apply this
approach at scale remains.
To better understand the barriers from the facility manager perspective, Gheisari and
Irizarry (2016) conducted a series of interviews, concluding that “start-up initial costs,
administrative costs and lack of BIM models of the facilities (they maintain) are the main
barriers of BIM adoption in (their) facility management practice”.

2.3 Historical barriers to adoption

The barriers noted above align with the broader literature for FM-BIM. A comprehensive
review by Volk et al. (2014) summarized the barriers above as:
 identifying the critical information required, the high level of effort to;
 develop;
 maintain the BIM; and
 information exchange between the BIM and FM systems.

The first barrier – to properly identify the information required to improve building
operational performance – is critical. While there is no limit to information types that can be
incorporated, given the level of effort to populate and maintain a model with operational
data, it should be kept to the minimum necessary. This information may be geometric,
topological, semantic or a combination thereof, and may relate to any one or more of the
following for FM applications: maintenance, life cycle assessment, space management and
planning, maintaining occupant comfort, sustainability, health and safety, condition
assessment and renovations. An FM is impacted by three key factors: business
environment, buildings and facilities characteristics and external interventions/factors
(Pärn et al., 2017); thus, their needs can be expected to vary as these elements change.
Because there is no “one-size fits all” approach, stakeholder engagement and involvement is
key to realizing the value of BIM in operations (Bryde et al., 2013).
The second barrier is model creation effort: many consider a complete as-built facility
information necessary for FM-BIM and the associated cost to create this complex geometry
poses a significant barrier to its use (Volk et al., 2014). Significant research has been
underway to capture as-built geometry (Ilter and Ergen, 2015; Volk et al., 2014; Yang et al.,
2016) to address geometric information. While 3D laser scanning and scan-to-BIM
technologies have developed significantly in recent years (Laing et al., 2015; Wang et al.,
2015; Yoon et al., 2015), two issues remain: their outputs remain limited to visible elements,
and the resulting point cloud models are large and computationally expensive. As such they
do not yet resolve this issue, demonstrating the need to limit reliance on this as-built
geometry. It has been demonstrated that owners require lower levels of geometric data and
significantly higher attribute (semantic and topological; Burg and Mealy, 2012) during
operational stages, reinforcing the need for research on integrating this non-geometric
building information into FM-BIMs.
Third, frequent changes, modifications and improvements made over time make it
difficult and costly to both create and maintain up-to-date information in a FM-BIM
(Volk et al., 2014; Lin and Su, 2013). A semi-automated approach has been developed for Lean-Agile
identifying and populating data in the FM context and tested in an initial case study (Pärn FM-BIM
et al., 2016), which noted the need for a means of automated information transfer with CAFM
systems. While several non-proprietary data exchange formats have been developed, notably
COBie (East, 2007), green building eXtensible Mark-up Language (Carmelsoft, 2017) and
industry foundation classes (buildingSMART, 2013), many CAFM systems are not
compatible with these formats, require the purchase of additional bridging software or – in
the case of many legacy systems – were not designed to communicate with third-party 679
software. The heterogeneous nature of such data sources and formats across CAFM systems
complicates data mining and transfer, (Kang and Choi, 2018) and hard-coded mapping
functions are not easily extended. Interoperability challenges thus pose a fourth barrier to
adoption and effective system integration is necessary for successful delivery of BIM projects
(Watson, 2011), particularly when multiple CAFM systems are in place.

3. Lean and Agile philosophies

The Lean philosophy was first proposed by Koskela (1992) in the manufacturing context,
and its application in project management focuses on reducing waste and operational
efficiency by restructuring project phasing and avoiding tasks until they are necessary
(Ballard and Howell, 2003). There has been significant research to adapt BIM to support
Lean projects (for example Arayici et al., 2011; Sacks et al., 2010), and recent research
(Rasmussen et al., 2017) applies this philosophy to BIM itself, developing the minimal
(simplest) models required to achieve the desired outcomes. The Agile philosophy arose
from computer science and prioritizes simplicity, collaboration, regular feedback,
adaptability to change and a focus on working software (outputs) over comprehensive
documentation (Beck et al., 2001). Like Lean, it promotes continual and incremental
improvement and seeks to eliminate or minimize non-value-adding activities. Within Agile,
“sprints” are short iterations to develop these incremental improvements and provide end-
users with working products in a minimum timeframe, allowing feedback and evaluation
after a minimum of investment.

4. Research methodology
The methodology used to develop the Lean-Agile approach itself followed Lean and Agile
processes, as illustrated in Figure 1, whereby a series of iterations were used to create and
continually refine the process based on lessons learned from the case studies. To start this
process, a literature review was undertaken to identify the state of FM-BIM adoption,
specific barriers and obstacles, review previous case studies, review BIM execution plans
(BxPs) to identify best practices and investigate Agile and Lean BIM approaches as they
have applied to design and process development. Next, the findings of this review were
synthesized into a conceptual approach for Agile FM-BIM implementation for FM-BIM.
This approach was then tested and refined through a series of iterations, each consisting of a
case study implementation and its evaluation. Within each case study, semi-structured
interviews with the FM/end-user(s) evaluated the effectiveness of the FM-BIM use(s)
presented at each sprint and determined information requirements for the next iteration.
This feedback was used to refine the resultant Lean-Agile process, as discussed in Section 4.
The first case study developed an FM-BIM for Kerr Hall East (KHE), a 42,000 m2
quadrant of a mixed-use academic (classroom/lab/office) building. This building was
selected as a worst-case demonstration project owing to its complexity, lack of existing
information and significant number of poorly documented renovations since its construction
in 1970. As such, the FM director felt that if FM-BIM were successfully implemented here, it
F
36,13/14

680

Figure 1.
Methodology

would be feasible across the campus. Four sprint cycles were undertaken in 2014/15 as Case
Study 1A, using a LOD 200 (AIA, 2013) base model supplemented by semantic information.
Staff turnover in 2015 resulted in new end-users and two more sprints added new types of
BIM uses (Case Study 1B). Automated data transfer protocols (Khaja et al., 2016) were
developed to streamline model maintenance starting from this point. The most valuable BIM
uses from Case Study 1 were integrated in a second case study for a new building, the
13,000 m2 (140,000 sf) Student Learning Centre (SLC) library extension, along with BAS data
mapping. The base model in this instance reflected the design coordination model available.
Scaling FM-BIM to the full campus (Case Study 3) required a critical analysis of base model
requirements. It was at this point that the Lean information schema was formalized,
allowing for automated building generation (Bortoluzzi et al., 2017), which was integrated
from this point forward.
The fourth case study was undertaken to provide the Dean of Engineering with
enhanced space management information on the 13 buildings (total 23,500 m2) hosting
faculty offices and laboratories, and is entering its second iteration. The final case study
addressed a new end-user type (hospital network) and modelled a 16-storey, 24,570 m2
paramedical training facility to form the basis for an entirely BIM-based FM system and has Lean-Agile
implemented its first BIM use. FM-BIM
All case studies were implemented using Autodesk Revit and outputs were shared as
both Revit files (models) and spreadsheets (room data schedules). The end-users for Case
Studies 1, 2 and 3 used Archibus as a CAFM system while the latter two studies’ end-users
had no CAFM system. For all case studies, the researchers implementing each use case
estimated time spent to the nearest hour. These hours represent the time specifically related
681
to use case development and exclude non-CAFM data collection and initial base model
creation (typically 15-30 hours/building).

5. The Lean-Agile facility management-based building information model

approach
The Lean-Agile approach arose as a combination of two project philosophies developed to
optimize project delivery in a context of continuous change. A standard Agile sprint
approach based on the scrum life-cycle developed by Golfarelli et al. (2013), adapted to suit
BIM creation steps identified in the BxP review. Applying Lean principles, this process
(Figure 2) was implemented by engaging the end-users (facility teams) as collaborators in
FM-BIM development and has been continuously refined based on constant evaluation
(discussed in Section 5.1).
This approach addresses identified barriers, resulting in flexible and expandable BIM
with increasing functionality over time, developed through close interaction with end-users.
Each iteration incrementally develops BIM content and CAFM data transfer interfaces

Figure 2.
Agile sprints for FM-
BIM development
F necessary to achieve the priority FM use functionality, minimizing modeling effort by using
36,13/14 non-geometric information in lieu of as-built geometry wherever possible and automating
key information exchanges to minimize maintenance costs. The incremental nature of this
approach provides flexibility and adaptability to both changing FM needs and information
availability. The Lean aspect of this approach is evident in:
 information management schema to avoid low-value BIM content;
682  data transfer solutions to overcome interoperability issues and reduce the long-term
BIM maintenance costs; and
 the built-in opportunity for continuous process improvement.

5.1 Develop base model

Where a full as-built BIM is not available, the first step is to create a base model to provide a
starting point for FM-BIMs and consists of stacked building floorplans with defined room
and level elements to permit mapping of CAFM data. The automated creation of these lean
models and how this process relates to the barriers to FM-BIM adoption have developed in a
previous study (Bortoluzzi et al., 2017). The entity–relationship diagram for this model
presents the minimal geometry necessary to implement the BIM uses from all case studies,
as illustrated in Figure 3, as formalized to support automated FM-BIM base model creation.
Once the base model has been developed, the next five stages – prioritize, identify, collect,
integrate and evaluate – are repeated for iterations of the process to add functionality to
support required FM applications.

5.2 Prioritize application

This step uses input from the facilities management team to identify and prioritize the next
most important FM activity to be supported by the BIM. This may be a new use or a

Figure 3.
Entity–relationship
diagram for base
model
refinement of a previously implemented use that is inadequate in its current form. This will Lean-Agile
vary from end-user to end-user, and the repetition of this stage at iterations provides FM-BIM
flexibility for the facilities team to shuffle their priorities and add new ones as the BIM is
enhanced and new functionality is demonstrated, rather than making a one-time list for
implementation. In Case Study 1, five BIM uses were sequentially identified (one per
iteration) by various FM team members as having high priority: space management and
tracking, lighting information to support retrofit feasibility studies, as-built drawing
integration, elevator inspection record and urgent issue visualization and maintenance 683
request clustering with a focus on thermal and odor complaints. After a change of staffing at
the end of Case Study 1A, the new facility engineer noted the following, which guided both
BIM uses in Case Study 1B:
We have 100,000 documents we have to deal with every year; if you can help us to reduce the
number we need to look at, you will get our attention.
As the BIMs were developed and evaluated, the FM team became more engaged, identifying
new possible ways to support their operational activities with this model and eventually
championing this research beyond the immediate team.

5.3 Identify relevant information

This step defines the necessary information and semantics to support the priority activity.
Table I summarizes the information criteria for each case study. The criteria to define the
necessary information for FM-BIM iterations draw extensively from recognized industry
best practices. A comprehensive review of the twelve most influential BxPs (McArthur and
Sun, 2015) noted a number of consistent recommendations, namely, to involve the owner and
O&M partner in defining required outputs; start the planning process early, particularly
regarding BIM use identification and selection; regularly review and update the BxP; and
document what is necessary to include in the model, including each element’s level of
development at each phase.
On the basis of this review, McArthur and Sun (2015) synthesized the following
recommendations of relevance to projects involving an FM partner:
 Project deliverables: Define the project deliverables and key metrics to define project
success including links to compliance documents and project requirements.
 Project goals: Define project goals during each project phase.
 BIM and facility data requirements: Identify third-party data required for BIM
(inputs) and end-user (owner/O&M partner) data management systems (outputs) to
ensure compatibility.
 BIM information exchanges: Define level of detail for each component at handover
and any information exchange standards and required interfaces with third party
software.
 BIM uses: Identify the most beneficial BIM uses for each phase and additional data/
information and BIM LOD required to implementing each one.

Applying Penn State (CIC, 2011) guidance to define these elements and required end-user
input, these recommendations can be repackaged into five questions:
Q1. What is the scope of activities the BIM must support?
Q2. What information is currently maintained in the existing FM systems, how is it
updated and what is its current format?
 F

684

Table I.
36,13/14

case studies
BIM information
criteria developed in
FM-BIM
Uses Require outputs and
(Case Study) Scope of activities Existing FM information Existing uses visualizations Granularity required

Space Generation of government Room #, room name, GFA, room ArchiBus Government space reports; input Data required at the
management space reports; update of type, occupancy, government outputs are for room signage; input to space room level with
(1A, 2, 3, 4) campus maps and building classification codes; manually inputs for planning; color-coded floorplans summaries at the floor,
signage updated in ArchiBus on 3-year triennial (by type, department, etc.) building, and campus
cycle; can export as .csv space reports levels
Lighting Identification and feasibility of No information currently N/A Cost and energy saving Building level
retrofits (1A) beneficial retrofits maintained estimates (peak and annual total) quantities
As-built Renovations and field repairs Selected (2 per cent of total) Reference for As-built geometry visualized in Incorporate all
drawing refer to existing as-built scans (.PDF) of original as-built future BIM model information from
integration drawings to provide locations blueprints; full as-built for work renovations available as-built
(1A) and quantities for equipment since 2010 and service drawings
and building services calls
Inspection Schedule elevator inspections; Previous elevator inspection Reports Summary of elevator inspection Data required at the
record review elevator inspection reports in PDF format reviewed by report data; elevators requiring elevator level; building
management reports to identify items facility inspections (>11 months since and full-campus
(1B, 2) requiring action engineer for previous inspection); summaries desired
actions visualization of elevators
required and requiring action
filed
Maintenance Review work orders and assign ArchiBus database (25 k-30 k Reviewed by Visualization of work order Each work order
and trades. ideally, identify priority work orders/year since May facility density on floorplans to support assigned to a room in
complaint areas 2010), for each time/date, engineer to prioritization the existing system
management location, problem type, assign
(1B, 2) description (unstructured text), appropriate
status (open/closed), dates trade(s)
opened and closed. (Unused
fields omitted)
(continued)
FM-BIM
Uses Require outputs and
(Case Study) Scope of activities Existing FM information Existing uses visualizations Granularity required

Building Visualization of alarms, trends 4,287 individual points (2,229 BAS Temperature visualization BAS points mapped to
automation and system operation at the equipment, 2,058 room sensors) monitored at floorplans to cross-reference room or equipment.
system equipment level; visualization from Siemens system with workstations. with occupant complaints. Equipment size and
(BAS) (2) of actual temperatures or set- minimal (<100 points) trending Alarms Visualization of changing room location based on
point deviations for 90 days. Update at change of trigger work conditions over time once coordinated
value only. Time-stamped point orders gateway device installed to construction model
dumps in .csv format possible stream data to BIM
Space Space planning and strategic List of laboratories and “This is a Color-coded floor plans of actual/ Room-level data with
utilization planning for future space associated researchers significant recommended space ratio dept. summaries
(4) allocations; (leased/new-build) maintained by the research office need” distributions and summary
annually reports
Lab Post-occupancy evaluation of No information is currently N/A Scoring of laboratory usefulness Room-level data
functionality the usefulness of laboratory maintained. A survey will be by evaluation characteristics;
(planned in space to identify priority developed by the end-users to visualization of trends across
4) renovations, new requirements gather the user input to populate buildings
and obsolete spaces this data
Space Understanding the total area Very little is maintained and “This is a Visualizations (color-coded Room-level data with
utilization (5) provided for each department, what exists is decentralized, e.g. problem” floorplans) by space type, dept. and building
space type and classification. staff phone list with office occupant type, number of summaries
Understand what space is numbers, room signage list occupants, room type,
being used for with in the (updated for renovations or staff department and classification
building changes), list of departments,
drawings
Lean-Agile
FM-BIM

685

Table I.
F Q3. How is this information currently used?
36,13/14 Q4. What are the required outputs and/or visualizations for the BIM application?
Q5. What level(s) of granularity are required for this information? Table II summarizes
the case study responses for each question and BIM use.
Giminez et al. (2016) argued that to be successful, a model must contain geometric (3D
686 element representation), topological (relationships between these elements) and semantic
(attribute) information in proportion to the intended end use. Because topological and
semantic information can often be mapped using parameters applied to existing objects,
these were prioritized. New geometric content was created where necessary to either host
the necessary semantic and topological information to support the FM activity or to visualize
this information effectively. Once the most efficient means to include the identified
information has been identified, the necessary parameter(s) are developed, with special
attention paid to parameter type and units to ensure maximum flexibility in future use.
Early case studies found that geometric content was the costliest to create, while
semantic and topological information supported BIM activities more efficiently. The
evolution of the proportion of model information content over the case studies is illustrated
as estimated and relative values based on the number of new elements (geometry) or
parameters (semantic) required to incorporating the use case in Figure 4. Note that
topological data are intrinsic to the BIM, and thus, topological values are hypothetical
estimates related to geometry. It is notable that after Case Study 2, the as-built information
was integrated as referenced drawings and become part of the base model. In Case Studies 4
and 5, the end-users required only semantic information, and thus, only minimal geometry
(room boundary polyhedron definitions) was necessary.

5.4 Collect data

In this stage, the relevant FM data are collected from the appropriate source(s), which may
include live CAFM system information, drawings or other electronic (or scanned)
documents. For example, space management data for Case Studies 1A, 2 and 3 were
obtained by exporting a .csv file from Archibus, the CAFM system used for space
management. Given the breadth of potential inputs to the BIM, pre-processing or data
mining is often necessary to restructure these into a format compatible with the BIM
information schema; where this is necessary, algorithm requirements are developed at this
stage. Electronic documents are filed in a new subfolder to facilitate linking to the BIM.

5.5 Integrate with computer-aided facilities management

This stage consists of three primary activities: implementing any required data pre-
processing algorithm(s); mapping the data to the BIM, including implementing information
exchange algorithms or automated scripts to update the model; and creating new
visualizations or schedules to support the priority BIM use. Previous work by Khaja et al.
(2016) demonstrated a means for automating the data mapping and transfer using
parametric scripts (dynamo) and macros; a modification of this approach is shown in Figure
5. This approach was used starting in Case Study 1B to automatically populate FM-BIM
parameters with updated data. Once data update is complete, the new use-case is then
integrated into an updated FM dashboard or other desired interface for testing with
available data or data streams before demonstration and evaluation.
 Base Model Use Case - Application
Lean-Agile
S S S S
FM-BIM
Case - #1A

687
G T G T G T G T
Space Management As-Built Integration
S S
Case - #1B

G T G T
Inspection Records Maintenance Records
S S S S
Case - #2

G T G T G T G Building Automation T
Space Management Maintenance Records System
S S
Case - #3

G T G T
Space Management
S S S S
Case - #4

G T G T G T G T
Space Management Space Utilisation ########
S S

Figure 4.
Case - #5

Conceptual
illustration of
S - Semantic evolving information
T - Topological
G - Geometric content by type in the
G T G
Space Management
T FM-BIM
F
36,13/14

688

Figure 5.
Data transfer
strategy

5.6 Evaluate
At the “Evaluate” stage, a demonstration of the new BIM functionality is made to the FM
team, who assess its success in achieving the defined goals and its value for full-scale
implementation through the following series of questions:
(1) Does the new BIM functionality add (potential) value? If so, what is the benefit?
(2) Does the BIM fully achieve the defined requirements?
 If so, is any of the information or effort in the current implementation
extraneous/unnecessary?
 If not, why? Is there information missing or does it need to be integrated
differently?
Successful iterations both add value and fully achieve the requirements. For example, the
Case 1a space management application achieved the desired functionality and provided
significant value: The space manager estimated that integration with the BIM would save
1,800 person-hours annually for space updates and reporting, compared with the current
practice of comparing and retracing (as necessary) the updated floor plans for campus
buildings as Archibus is unable to directly read the campus master BIM and CAD files. To
ensure this implementation was Lean, Question 1a was asked after implementation and
extraneous elements or processes are removed or modified before further implementation.
When implemented in Case Study 2, the automation process had been improved to a two-
step process, while a refined Python mapping script optimized the space mapping,
permitting a single data file to be used to update the whole-campus model in Case Study 3 Lean-Agile
implementation. Similarly, the work order mapping initially used individual parameters for FM-BIM
each historical data point and was refined to a single list parameter accessed through a
customized visualization tool to scroll through these time series.
If a BIM use is unsuccessful but offers value, Question 1b informs modifications in a
second iteration. In Case Study 4, the initial space utilization implementation was met with a
positive response, both regarding the usefulness of the visualizations and the timeliness of
delivery: “To have this picture is extremely valuable [. . .] and will help in (upcoming) 689
discussions”. At the same time, some of the space utilization requirements were not
identified in the initial meeting but were apparent during the evaluate phase. A second
iteration addressed these newly identified requirements and prompted a larger-scale data
collection exercise to address this BIM use. Similarly, these newly identified requirements
allowed the end-users to better define the third BIM use (lab functionality) and demonstrates
the value of Agile in developing the BIM; by engaging the end-user regularly to provide
feedback on incremental improvements, new applications and BIM use requirements can be
more quickly identified and deployed.
Similarly, there were cases when successful subsequent iterations can further enhance
this functionality while the partially successful BIM use can be deployed. For example,
when the first iteration of the maintenance and complaint management use case was
presented to the facility engineer, he responded, “This is helpful – it prompts us to ask the
right questions”, adding that incorporating Building Automation System (BAS) data would
add further value in assessing work orders, thus informing a future iteration after that BIM
use was developed. Similarly, the dean noted that “the picture provided is extremely helpful
and timely” in response to the initial space utilization prototype, but noted additional data
required as well as errors in the centralized space management database requiring
correction.
Demonstrated BIM uses that do not offer value – whether successful– are not
implemented. For example, the lighting retrofit feasibility application in Case Study 1 did
not provide value to the operations team, who felt that consultants would provide such
service as part of their scope and was abandoned. Because the Lean approach minimized
effort, the resultant loss was minimal (20 person-hours), demonstrating the value of low-cost
“sprints” to demonstrate initial functionality and reduce the financial investment and
associated risk in new BIM use development.

6. Discussion
The presented approach was evaluated after each sprint. A meta-evaluation is presented
below based on five case studies, identifying strengths, limitations and refinement
opportunities.

6.1 Evaluation of proposed approach

While questions used in the evaluate stage provide insight on iteration success, the overall
approach was evaluated based on the following categories and questions, developed to align
with Koskela (1992) and Agile (Beck et al., 2001) principles and to address noted research
gaps:
 Success: based on the evaluate stage feedback, how successful was each case
study?
 Flexibility: was the approach agile and adequate to address the breadth of FM
applications considered?
F  Efficiency: was the approach lean? What resources were required to implement each
36,13/14 application? Where inefficiencies were identified in early implementation, did the
process encourage/support optimization?
 Continuous improvement: What are potential improvements to the process?

As noted in the case study, the majority of BIM uses were successful; the cost associated
690 with unsuccessful (low value) was insignificant because they were abandoned after the first
iteration and thus required minimal resources.
The approach demonstrated flexibility in both the diversity of BIM uses implemented
and in permitting the shift in type of application. Case Study 1A focused strictly on data
mapping to the BIM while Case Studies 1B and 2 synthesized key information during post-
processing. Case Study 3 provided significant insight on how to streamline and automate
the process, which dramatically decreased the time required to develop base models for the
following case studies.
This process was found to be highly efficient. The information criteria successfully
informed necessary parameter development and sprints generally achieved the desired FM
outputs in all cases, as discussed in the previous section. The refined Lean-Agile approach
allows the FM team to evaluate new BIM uses after between 1 and an average of 15 person-
hours of effort (excludes base model creation and non-CAFM field data collection as well as
any custom dashboard visualizations developed), as presented in Table II. This limited
investment on low-value applications while permitting the rapid deployment of high-value
applications.
The low-geometry approach affected how information was visualized in the BIM but did
not detract from functionality. As-built drawings and semantic information were visible in
2D, providing the known building geometry and access to all non-geometric information. In
3D, the lack of geometry is immediately apparent because the base geometry is effectively
invisible, providing a wireframe of related spaces and volumes. Geometric elements are
added as required to support new applications; for example, envelope components facilitated
campus visualization and energy analysis in Case Study 3. The impact of this low-geometry
approach on information accessibility was considered from two perspectives: the
accessibility of the information in the BIM and challenges in implementing required
visualizations. In some ways, the data were more accessible in the BIM than in original
sources, which were stored across multiple software platforms across the department. by
using shared parameters to capture information, schedules were easily generated and could
be manipulated in-situ or exported in spreadsheet format for wider distribution.
Visualization development was simplified by inbuilt functionality such as display filters to
change element and room colors based on parameter values, and these views were laid out
on a sheet and exported to PDF to facilitate dissemination. Key challenges in visualization
arose with determining how to map information related to a conceptual object rather than a
physical one. For example, floor level summaries had to be mapped to the appropriate slabs.
Early applications often created more parameters than necessary to map relevant data to
the BIM and, thus, were relatively inefficient. As inherent BIM capabilities were explored
and leveraged on earlier applications, newer applications could be more efficiently
developed. This suggests that while the process itself is flexible and straight-forward, there
was a learning curve necessary to develop applications efficiently.
The overall success of this approach is best captured by the following feedback from the
facilities engineer, who was initially skeptical but eventually served as a champion for this
project: “It has always been easy to support (Case Study 2); I truly believe in the merit of
your work”. This led to additional case studies on campus buildings, interest in expanding
 BIM uses and
Lean-Agile
time for Continuous FM-BIM
Case implementation Success (per Flexibility Efficiency improvement
Study (hours) FM) (“Agile”) (“Lean”) (s) identified Comments

1A Space Moderate: High – Low – base Need for All parameter

management (15), two FM approach model more automated data manually
lighting retrofit team did not detailed than process to input 691
(5), as-built members require required; too bulk import
integration (<1) positive, modification much manual parameters
limited by use case input
adoption by
others
1B Inspection record High: good High Moderate: Need to 90 per cent of
management (20), feedback inspection overcome time was spent
maintenance and from FM record macros Revit developing
complaint stakeholders effective; ordering of macros and
mapping (40) maintenance elements and mapping
macros need order on algorithms
optimization room ID
2 Inspection record High: good High Moderate: Need for BAS use driven
management (2), feedback macros very batch- by number of
maintenance and from broader effective for; generation new
complaint FM team manual and mapping parameters
mapping (2), BAS parameter of new required for the
integration (100) creation slow parameters 4,500 mapped
points
3 Space High: broad High High: no Data Inefficient
management (9; 1- adoption and unused mapping parameter
5 minutes/storey) support information; script used mapping script
full automation inefficient
sorting
algorithm
4 Space utilization High: fully High High N/A Optimized
(5) achieved mapping script
desired goals
5 Space High: High High N/A Parameter
management (2) immediate mapping took
adoption by 12s/storey;
FM team in remaining time
lieu of was Table II.
commercial visualization Case study
CAFM creation evaluation summary

to include leased off-campus spaces such as laboratories and research centers and an
introduction to the hospital network (Case Study 5).

6.2 Information management structure

The information management structure presented addresses two key barriers to adoption: it
identifies the necessary information to host in the FM-BIM and reduces the level of effort
necessary for its creation. The schema also supports the automated data transfer from the
CAFM to the BIM, indirectly addressing the barriers of model maintenance cost and
interoperability. This structure provides flexibility to define the necessary information for
F both hard- and soft-FM-BIM uses and the means to map the full range of information types
36,13/14 onto the base model, while limiting investment in detailed geometric modelling to that
required for priority end-uses.
From Case Study 3 onwards, room elements defined from boundary elements overlaid
with the floor plans were found adequate to map and visualize CAFM data at the room level
using color-coded overlays. Building-level information can be assigned to symbolic elements
692 such as signage to permit visualization and ease of retrieval. Equipment and other 3D-
modeled elements were created only where data-hosting required it and were represented
using the most appropriate generic equipment family, with parameters added and revised to
incorporate the required system and semantic information.

6.3 Data transfer

The automation of data transfer from CAFM systems and other data sources addresses
barriers of ongoing model maintenance cost and interoperability. Pre-processing is of critical
importance in this process to reformat the data and align it with the defined parameters and,
thus, streamlines this automated update of BIM information. To protect the integrity of the
CAFM systems, the proposed approach clones this data within the BIM, providing two-way
write functionality only where specifically desired to the FM.
The integration of Big Data tools and sensor networks with FM-BIM is an area of
significant potential and a recommended area of future research. A pilot project currently
underway is exploring data architectures to stream BAS and other CAFM data live onto a
private cloud. This will provide a robust platform for automating the data pre-processing
and integrating analytics before BIM integration, providing significant visibility into the
near real-time operation of FM assets.

7. Conclusions
The proposed conceptual approach to BIM provides a means to rapidly implement and
evaluate FM applications with the minimum of modelling necessary to achieve the required
functionality.
Practical implications: Because FM applications are developed on an as-required basis,
minimal effort is required before the business case for full deployment can be established.
The low-geometry approach circumvents the need for fully modelled 3D geometry at the
outset, permitting lower-cost but high-value uses to be implemented before those requiring
more significant investment (e.g. 3D as-built creation). This approach thus provides an
alternate means for facilities staff to leverage BIM to visualize CAFM data and addresses
the cost barrier to adoption. In addition, 2D to 3D automation research (Giminez et al., 2016;
Bortoluzzi et al., 2017) has the potential to further reduce this barrier.
Research implications: The focus on semantic, rather than geometric, information and its
inherent topology has implications for the development of FM-BIM ontology. Unlike
construction, where physical characteristics of all elements are critical, this ontology would
prioritize operational and relational characteristics, incorporating necessary physical data to
support asset management and maintenance activities. Pärn et al. (2016) provided an initial
totem toward this ontology, which is a planned topic for future study.
Limitations: There are two limitations to the current research. First, the number of
implemented BIM uses in the demonstration project is limited, and while it represents a
breadth of possible applications, it is by no means comprehensive. The FM-BIMs will be
extended in both scope and degree of BIM use implementation in future case studies to
address this issue. Second, as with BxP guides, the proposed approach is non-prescriptive,
requiring engagement and significant input from the FM as well as careful consideration of
required parameters to develop incremental BIM functionality. The development a Lean Lean-Agile
ontology is expected to provide specific guidelines and simplify parameter identification and FM-BIM
development.
Concurrent and future research: Current work builds upon the existing case studies; for
example, the BAS mapping BIM use is being adapted to integrate near real-time streaming
of data along with analytics and 4D visualization of system performance. This contributes
to a large FM platform development project, including a novel software architecture and
data ingestion system (Fokaefs et al., 2017) and the integration of machine learning and 693
predictive control. New BIM analytics and the Lean ontology, building upon the schema
presented, are also planned areas of research.

References
AIA (2013), AIA Document G202–2013, Project Building Information Modeling Protocol Form,
American Institute of Architects, Washington, DC.
Arayici, Y., Coates, P., Koskela, L., Kagioglou, M., Usher, C. and O’Reilly, K. (2011), “Technology
adoption in the BIM implementation for lean architectural practice”, Automation in
Construction, Vol. 20 No. 2, pp. 189-195.
Arayici, Y., Onyenobi, T. and Egbu, C. (2012), “Building information modelling (BIM) for facilities
management (FM): the mediacity case study approach”, International Journal of 3-D
Information Modeling, Vol. 1 No. 1, pp. 55-73.
Ballard, G. and Howell, G. (2003), “Lean project management”, Building Research and Information,
Vol. 31 No. 2, pp. 119-133.
Becerik-Gerber, B., Jazizadeh, F., Li, N. and Calis, G. (2012), “Application areas and data requirements
for BIM-Enabled facilities management”, Journal of Construction Engineering and Management,
Vol. 138 No. 3, pp. 431-442.
Beck, K., Beedle, M., Van Bennekum, A., Cockburn, A., Cunningham, W., Fowler, M., Grenning, J.,
Highsmith, J., Hunt, A., Jeffries, R. and Kern, J. (2001), “Manifesto for agile software
development”, available at: http://agilemanifesto.org (accessed 15 June 2015).
Bortoluzzi, B., Sobieraj, D. and McArthur, J. (2017), “Automating the creation of facility and energy
management building information models”, Lean and Computation in Construction Congress,
Heraklion, pp. 153-160.
Brooks, T. and Lucas, J. (2014), “A study to support BIM turnover to facility managers for use after
construction”, Computing in Civil and Building Engineering, pp. 243-250.
Bryde, D., Broquetas, M. and Volm, J.M. (2013), “The project benefits of building information modelling
(BIM)”, International Journal of Project Managmeent, Vol. 31 No. 7, pp. 971-980.
buildingSMART (2013), IFC4: The New BuildingSMART Standard, buildingSMART International,
Northamptonshire.
Burg, B. and Mealy, C. (2012), Xavier University Facility Management BIM Integration – a Case Study,
buildingSMART, Dallas.
Carmelsoft (2017), “Software list”, available at: www.gbxml.org/index.html (accessed 16 September
2017).
Chen, K., Lu, W., Peng, Y., Rowlinson, S. and Huang, G.Q. (2015), “Bridging BIM and building: from a
literature review to an integrated conceptual framework”, International Journal of Project
Management, Vol. 33 No. 6, pp. 1405-1416.
CIC (2011), BIM Project Execution Planning Guide – Version 2.1, Computer Integrated Construction
(CIC) Research Program PA State University, University Park, PA.
Eadie, R. et al. (2013), “BIM implementation throughout the UK construction project lifecycle: an
analysis”, Automation in Construction, Vol. 36, pp. 145-151.
F East, W.E. (2007), Construction Operations Building Information Exchange (COBie), Engineer Research
and Development Center Construction Engineering Research Lab, Champaign, IL.
36,13/14
Eastman, C., Teicholz, P., Sacks, R. and Liston, K. (2011), BIM Handbook: A Guide to Building
Information Modeling for Owners, Managers, Designers, Engineers and Contractors, 2nd ed.,
Wiley, Hoboken.
Fokaefs, M. et al. (2017), Leveraging Existing Sensor Networks as IoT Devices for Smart Buildings,
Springer, Barcelona.
694
Gheisari, M. and Irizarry, J. (2016), “Investigating human and technological requirements for successful
implementation of a BIM-based mobile augmented reality environment in facility management
practices”, Facilities, Vol. 34 Nos 1/2, pp. 69-84.
Giminez, L., Robert, S., Suard, F. and Zreik, K. (2016), “Automatic reconstruction of 3D building models
from scanned 2D floor plans”, Automation in Construction, Vol. 63, pp. 48-56.
Golfarelli, M., Rizzi, S. and Turricchia, E. (2013), “Multi-sprint planning and smooth replanning: an
optimization model”, Journal of Systems and Software, Vol. 86 No. 9, pp. 2357-2370.
Hanna, A., Yuetter, M. and Aoun, D. (2013), “State of practice of building information modeling in the
electrical construction industry”, Journal of Construction Engineering and Management,
Vol. 140 No. 12, pp. 05014011-1-11.
Ilter, D. and Ergen, E. (2015), “BIM for building refurbishment and maintenance: current status and
research directions”, Structural Survey, Vol. 33 No. 3, pp. 228-256.
Kang, T. and Choi, H. (2018), “BIM-based data mining method considering data integration and
function extension”, KSCE Journal of Civil Engineering, Vol. 22 No. 5, pp. 1523-1534.
Kassem, M., Kelly, G., Dawood, N., Serginson, M. and Lockley, S. (2015), “BIM in facilities management
applications: a case study of a large university complex”, Built Environment Project and Asset
Management, Vol. 5 No. 3.
Khaja, M., Seo, J.D. and McArthur, J.J. (2016), Optimizing BIM Metadata Manipulation Using
Parametric Tools, Elsevier, Tempe, AZ, pp. 259-266.
Kiviniemi, A. and Codinhoto, R. (2014), “Challenges in the implementation of BIM for FM —
case Manchester town hall complex”, Computing in Civil and Building Engineering,
pp. 665-672.
Kivits, R.A. and Furneaux, C. (2013), “BIM: enabling sustainability and asset management through
knowledge management”, The Scientific World Journal, Vol. 2013.
Koskela, L. (1992), Application of the New Production Philosophy, Center for Integrated Facility
Engineering, Stanford.
Kreider, R., Messner, J. and Dubler, C. (2010), Determining the Frequency and Impact of Applying BIM
for Different Purposes on Building Projects, Loughborough University, Loughborough.
Laing, R., Leon, M., Isaacs, J. and Georgiev, D. (2015), Scan to BIM: The Development of a Clear
Workflow for the Incorporation of Point Clouds within a BIM Environment, WIT Press, Bristol,
pp. 297-307.
Lin, Y.-C. and Su, Y.C. (2013), “Developing mobile- and BIM-Based integrated visual facility
maintenance management system”, The Scientific World Journal, Vol. 2013, p. 124249.
Love, P. et al. (2014), “A benefits realization management building information modeling framework for
asset owners”, Automation in Construction, Vol. 37, pp. 1-10.
McArthur, J. and Sun, X. (2015), Best Practices for BIM Execution Plan Development for a P3 (PFI)
Design-Build-Finance-Operate-Maintain Project, WIT Press, Bristol, pp. 119-127.
Pärn, E.A., Edwards, D.J. and Draper, R. (2016), A Case Study of Building Information Modelling
Enabled ‘Information Totem’ for Operations and Maintenance Integration, CIB, Tampere.
Pärn, E.A., Edwards, D.J. and Sing, M.C.P. (2017), “The building information modelling trajectory in
facilities management: a review”, Automation in Construction, Vol. 75, pp. 45-55.
Parsanezhad, P. and Dimyadi, J. (2013), Effective Facility Management and Operations via a BIM-Based Lean-Agile
Integrated Information System, Polyteknisk Boghandel og Forlag, Copenhagen, pp. 442-453.
FM-BIM
Rasmussen, M.H., Pauwels, P., Hviid, C.A. and Karlshoj, J. (2017), “Proposing a Central AEC ontology
that allows for domain specific extensions”, Lean and Computation in Construction Congress,
Heraklion.
Sacks, R., Koskela, L., Dave, B.A. and Owen, R. (2010), “Interaction of lean and building information
modeling in construction”, Journal of Construction Engineering and Management, Vol. 136 No. 9,
pp. 968-980. 695
Volk, R., Stengel, J. and Schultmann, F. (2014), “Building information modeling (BIM) for existing
buildings — literature review and future needs”, Automation in Construction, Vol. 38,
pp. 109-127.
Wang, C., Cho, Y.K. and Kim, C. (2015), “Automatic BIM component extraction from point clouds of
existing buildings for sustainability applications”, Automation in Construction, Vol. 56, pp. 1-13.
Watson, A. (2011), “Digital buildings — challenges and opportunities”, Advanced Engineering
Informatics, Vol. 25 No. 4, pp. 573-581.
Yang, J., Shi, Z.K. and Wu, Z.Y. (2016), “Towards automatic generation of as-built BIM: 3D building
facade modeling and material recognition from images”, International Journal of Automation
and Computing, Vol. 13 No. 4, pp. 338-349.
Yoon, S., Jung, J. and Heo, J. (2015), “Practical implementation of semi-automated as-built BIM creation
for complex indoor environments”, The International Archives of Photogrammetry, Remote
Sensing and Spatial Information Sciences, Vol. 40 No. 4, pp. 143-146.

Corresponding author
J. J. McArthur can be contacted at: [email protected]

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: [email protected]